crewAI · Architecture
Created · run_mwili5cqb9d827bd28eabbb3 · github.com/crewAIInc/crewAI @ main · commit 2817a951f308 · model claude-opus-5-5
How the CrewAI monorepo is built: six Python packages that let a developer define agents, tasks, crews and flows, run them against LLM providers and tools, and optionally deploy them to the CrewAI AMP platform.
CrewAI is an open-source Python framework for building teams of AI assistants that work together on a job. A developer describes a few agents (each with a role, a goal and the tools it may use), the tasks they must complete, and how they cooperate: as a crew that works through the tasks in order or under a manager, or as a flow that strings steps, crews and human approvals into a longer process. CrewAI then runs that team inside the developer's own program, sending each agent's thinking to a large language model and letting it call search engines, scrapers, databases and other services on the way. It is a library and a command-line tool, not a hosted service: nothing in this repository runs on a server of its own, and the optional hosted side, the CrewAI AMP platform, lives outside it.
The repository is a uv workspace of six packages, all at version 1.15.41: crewai, crewai-core, crewai-tools, crewai-files, crewai-cli and crewai-devtools. The crewai package holds agents, tasks, crews, flows, memory, knowledge and the LLM layer; crewai-core holds what the library and the command line share (authentication, settings, the platform client, telemetry); crewai-tools ships 84 tool directories; crewai-files handles files passed to models; crewai-cli is the crewai command, whose console script points at crewai_cli.cli:crewai and is declared by both the crewai and crewai-cli packages; crewai-devtools is the maintainers' release tooling. Supported Python is >=3.10,<3.14, and the product ships as PyPI packages.
| Actor | Who or what it is | How it meets the system |
|---|---|---|
| The developer | The engineer building an AI application on CrewAI. | Imports the library into their own code, and uses the crewai command to scaffold, run, train, test and deploy a project. The only actor who runs this repository's code directly. |
| The end user of the developer's app | Whoever the developer built the application for. | Never touches CrewAI directly; their requests arrive as the inputs a crew or flow is kicked off with, and their answers to a paused flow come back as human feedback. |
| The LLM provider | The model service that does the reasoning: seven native provider packages (anthropic, azure, bedrock, gemini, openai, openai_compatible, snowflake), with LiteLLM as an optional extra pinned >=1.105.3,<1.105.4 for the rest. | Called over the network on every agent step with the developer's own API key, read from the environment. |
| Third-party tool APIs | Search, scraping, browser, database and SaaS services the 84 tools wrap, plus MCP servers and remote A2A agents. | Called by an agent when it decides to use a tool, again with keys the developer supplies. |
| CrewAI AMP | The CrewAI platform at https://app.crewai.com by default, operated outside this repository. | Optional. The CLI logs in to it with the OAuth2 device flow, deploys crews to it, publishes tools and skills to it, and sends execution traces to it when tracing is enabled. Anonymous telemetry, on by default, goes separately to https://telemetry.crewai.com:4319. |
| Maintainers | The CrewAI team releasing the packages. | Use the five crewai-devtools console scripts and 19 GitHub Actions workflows to test, version and publish to PyPI. |
What the system is for, in one line: to let a developer turn "a few specialised assistants working through a job together" into ordinary Python that runs where their application runs, with the hosted platform as an optional place to deploy and watch it.
Good, at the low end. Standards are enforced mechanically in CI (ruff lint and format checks, strict mypy with the pydantic plugin, pytest with the network blocked under random order and timeouts, pip-audit and CodeQL), with 542 test files for 985 source files and clear package boundaries. It is held back by concentration rather than sloppiness: 104 source files over 500 lines with several over 3,500, 168 functions over 100 lines, four two-way pairs between source packages in the module graph (two of them real cycles, section 4), 126 empty error handlers and 22 verified redundant items in section 17. The technical assessment gives the reasoning per criterion.
| Rating | Scale | Score |
|---|---|---|
| Engineering standards | 7/10 Good | |
| Security | 7/10 Good | |
| Configuration and secrets | 6/10 Fair | |
| Data and persistence | 6/10 Fair | |
| Error handling and resilience | 7/10 Good | |
| Availability and recovery | 5/10 Fair | |
| Performance and scale | 6/10 Fair |
The process CrewAI exists to serve is a developer's journey from an idea ("a researcher and a writer who produce a weekly market brief") to a running, improving AI team, and optionally to a hosted deployment. The steps below are the ones a developer would recognise from the documentation and the crewai command, in the order they are usually taken. Steps 1 to 5 are local; steps 6 and 7 involve the CrewAI AMP platform and are optional.
Two things follow from this shape. The library is useful on its own: a developer who never logs in still completes steps 1 to 6 (except the platform grading in crewai eval). And the platform is reached only through explicit commands, apart from anonymous telemetry, which is on by default, and trace export, which is controlled by CREWAI_TRACING_ENABLED and crewai traces enable or disable.
Each step is reached through a command of the crewai CLI, declared in lib/cli/src/crewai_cli/cli.py. The handlers named in the steps are these.
| Step | Command | Handler | Where | What it does |
|---|---|---|---|---|
| Scaffold a project | crewai create | create | lib/cli/src/crewai_cli/cli.py:225 | Asks for the project type and name when they are missing, then scaffolds a crew, flow, agent, tool, skill or template project. For a new crew or flow in an interactive terminal it offers to start it at once in a fresh process running crewai run. |
| Run a crew or flow | crewai install | install | lib/cli/src/crewai_cli/cli.py:711 | Installs the project's dependencies with uv sync and exits with uv's status code when that fails. |
| Run a crew or flow | crewai run | run | lib/cli/src/crewai_cli/cli.py:754 | Runs the crew, flow or standalone agent configured for the project in the current directory, or a definition file passed with --definition. A standalone agent opens a terminal chat unless --inputs asks for one kickoff. |
| Review outputs and memory | crewai log-tasks-outputs | log_tasks_outputs | lib/cli/src/crewai_cli/cli.py:484 | Loads the task outputs stored from the latest crew kickoff and prints them, or says none were logged. |
| Review outputs and memory | crewai memory | memory | lib/cli/src/crewai_cli/cli.py:613 | Opens a terminal interface for browsing memory scopes and recalling stored memories, with an optional storage path and embedder settings. |
| Review outputs and memory | crewai checkpoint resume | checkpoint_resume | lib/cli/src/crewai_cli/cli.py:1593 | Records that a checkpoint was used and resumes the run from the named checkpoint, or the most recent one when none is named. |
| Improve the crew | crewai train | train | lib/cli/src/crewai_cli/cli.py:410 | Trains the crew for a number of iterations (5 by default), collecting human feedback into a trained-agents file (trained_agents_data.pkl by default). |
| Improve the crew | crewai test | test | lib/cli/src/crewai_cli/cli.py:685 | Runs the crew several times and has an evaluating model score the results. |
| Improve the crew | crewai replay | replay | lib/cli/src/crewai_cli/cli.py:457 | Re-runs the last crew execution starting from a chosen task, reusing the stored outputs of the tasks before it. |
| Improve the crew | crewai eval | eval_command | lib/cli/src/crewai_cli/cli.py:819 | Has CrewAI AMP grade the project's last traced run, running the crew first when there is none or the project changed since; with --models it compares models on the project's deployment. |
| Publish tools and skills | crewai tool publish | tool_publish | lib/cli/src/crewai_cli/cli.py:1103 | Logs in to the tool repository and publishes the tool project in the current directory, publicly or privately. |
| Publish tools and skills | crewai skill publish | skill_publish | lib/cli/src/crewai_cli/cli.py:1155 | Publishes the skill in the current directory, scoped to the developer's organisation. |
| Deploy to CrewAI AMP | crewai login | login | lib/cli/src/crewai_cli/cli.py:914 | Clears the stored user settings and signs the developer in to CrewAI AMP through the OAuth2 device flow. |
| Deploy to CrewAI AMP | crewai deploy create | deploy_create | lib/cli/src/crewai_cli/cli.py:959 | Creates a new deployment of the project on CrewAI AMP, validating the project first unless told to skip it. |
| Deploy to CrewAI AMP | crewai deploy push | deploy_push | lib/cli/src/crewai_cli/cli.py:993 | Pushes the current project to an existing deployment, identified by UUID or by the project. |
Work enters CrewAI in six ways, and only one of them is how a production system uses it. A developer's own application imports the library and calls a kickoff method on a Crew, Flow or Agent: that is the runtime path, and it runs in the caller's process. Everything else is tooling around it: the crewai command line that scaffolds, runs, tests and deploys projects; five maintainer console scripts from crewai-devtools; eight Python scripts run by hand or by CI; nineteen GitHub Actions workflows; and the browser event listeners of the flow visualisation page that crewai flow plot generates. The inventory below has 125 rows: 58 CLI commands, 5 devtools scripts, 8 script entry points, 17 library kickoff methods, 19 workflows and 18 browser listeners.
There are no HTTP routes, queue consumers or webhook receivers in the repository. The A2A integration ships server authentication schemes and an agent card builder, but no server: an application that wants to expose an agent over A2A mounts those pieces in its own web framework. The only schedulers are GitHub Actions cron triggers (nightly, stale, build-uv-cache, vulnerability-scan).
The crewai console script points at crewai_cli.cli:crewai and is declared by both the crewai and crewai-cli packages (lib/crewai/pyproject.toml lines 161 to 162, lib/cli/pyproject.toml lines 34 to 35), so installing either one puts the same command on the path. The crewai CLI has 58 invocable commands: 15 top-level commands plus 43 in 13 groups (eval, deploy, tool, skill, template, flow, triggers, org, enterprise, config, env, traces, checkpoint); eval and checkpoint also run when invoked bare. The org group is declared to run bare as well, but its body is empty, so crewai org on its own does nothing and is not counted. lib/cli/src/crewai_cli/cli.py line 1635 lets the module run as python -m crewai_cli.cli, which is how crewai create starts a run in a fresh process; it reaches the same commands. The old import path is kept as a shim: crewai.cli is a deprecated import shim that redirects crewai.cli.X to crewai_cli.X and emits a DeprecationWarning.
Most handlers are thin. Each imports its implementation lazily (the heavy modules load only when a command runs) and hands off to a command class or function in crewai_cli. The ones that run user code do it out of process: crewai run, install and related commands spawn uv run subprocesses inside the user's project (a project locked with Poetry gets poetry run instead). Commands that talk to CrewAI AMP go through the platform client and the stored login. Rows below give the command as typed, the line of the handler's def in lib/cli/src/crewai_cli/cli.py, and the handler.
| Command | Declared at | Handler | What it does |
|---|---|---|---|
crewai uv | cli.py:144 | uv | Runs any uv command in the current project with the credentials of every installed tool added to its environment. It needs a pyproject.toml and exits with uv's own code when uv fails. |
crewai create | cli.py:225 | create | Scaffolds a new crew, flow, agent, tool, skill or template project, prompting for the type and name when they are omitted (and refusing to prompt when CREWAI_DMN is set). At an interactive terminal it then offers to start the new crew or flow in a fresh process. |
crewai version | cli.py:370 | version | Prints the installed crewai version, and with --tools the installed crewai-tools version. |
crewai train | cli.py:410 | train | Backfills the project id, then trains the crew for n iterations (default 5) and saves each agent's learned suggestions to a pickle (default trained_agents_data.pkl). |
crewai replay | cli.py:457 | replay | Re-runs the crew from a given task id onward, reusing the stored outputs of the tasks before it, optionally with a trained-agents pickle. Errors are printed rather than raised. |
crewai log-tasks-outputs | cli.py:484 | log_tasks_outputs | Prints the task id, expected output and output of every task stored from the latest crew kickoff. |
crewai reset-memories | cli.py:541 | reset_memories | Deletes the selected stores: memory, knowledge, agent knowledge, the latest kickoff outputs, or all of them. The deprecated --long, --short and --entities flags are mapped to --memory. |
crewai memory | cli.py:613 | memory | Opens a Textual terminal UI over the LanceDB memory store (default ./.crewai/memory) to browse scopes and run recall queries with a chosen embedder. |
crewai test | cli.py:685 | test | Backfills the project id, runs the crew n times (default 3) and has an evaluation model score each run. |
crewai install | cli.py:711 | install | Locks and installs the project's dependencies with uv, passing any extra arguments through, and exits with uv's status when it fails so CI can key off it. |
crewai run | cli.py:754 | run | Runs the crew, flow or standalone agent configured in the project's [tool.crewai] table, or a declarative flow named by --definition, with optional JSON inputs. The run happens in a subprocess inside the project's own environment. |
crewai update | cli.py:898 | update | Rewrites a Poetry project's pyproject.toml for uv and keeps the original as pyproject-old.toml. It does not upgrade CrewAI. |
crewai login | cli.py:914 | login | Clears the saved user settings and signs in to CrewAI AMP through the OAuth2 device flow, storing the resulting token for later commands. |
crewai logout | cli.py:929 | logout | Deletes the stored tokens and user settings, or with --reset returns every CLI setting to its default. |
crewai chat | cli.py:1276 | chat | Starts a terminal conversation with the crew in which the chat LLM collects the crew's inputs and answers the user. |
crewai eval (bare) | cli.py:819 | eval_command | Sends the project's last traced run to CrewAI AMP for grading, running and tracing the crew first when there is no current run, and exits 0 only when the goal gate passes. With --models it instead compares models on the project's deployment. |
crewai eval export | cli.py:882 | eval_export_command | Downloads an evaluation's report, outputs, judge checks and traces into ./crewai-eval-ID/, with a .gitignore so the data is never committed. |
crewai deploy create | cli.py:959 | deploy_create | Backfills the project id, runs the pre-deploy checks and creates a deployment on CrewAI AMP from the project's git remote, or from an uploaded ZIP with --zip. |
crewai deploy list | cli.py:971 | deploy_list | Lists the deployments the logged-in user can see on CrewAI AMP. |
crewai deploy push | cli.py:993 | deploy_push | Backfills the project id, validates the project and redeploys it to an existing deployment (by --uuid or the project's own). AMP holds the push back while a model comparison is running unless --force is given. |
crewai deploy validate | cli.py:1005 | deploy_validate | Runs the same pre-deploy checks locally without contacting the platform and exits non-zero on any blocking issue. |
crewai deploy status | cli.py:1019 | deply_status | Fetches the status of a deployment from CrewAI AMP. The handler's name is misspelt in the source; the command name is not. |
crewai deploy logs | cli.py:1027 | deploy_logs | Fetches a deployment's logs from CrewAI AMP. |
crewai deploy remove | cli.py:1035 | deploy_remove | Deletes a deployment on CrewAI AMP. |
crewai tool add | cli.py:1048 | tool_add | Adds tool references (a built-in tool, platform:app or custom:name) to the standalone agent project in the current directory, validating platform applications on AMP first. |
crewai tool setup | cli.py:1061 | tool_setup | Validates the platform applications listed in the agent definition on AMP and saves the Enterprise Action Auth Token to the project's .env. |
crewai tool create | cli.py:1074 | tool_create | Deprecated alias of crewai create tool: warns, then scaffolds a custom tool project. |
crewai tool install | cli.py:1085 | tool_install | Logs in to the Tool Repository and installs a published tool by its handle into the current project. |
crewai tool publish | cli.py:1103 | tool_publish | Logs in and publishes the current tool project to the Tool Repository as public or private; --force skips the git remote checks. |
crewai skill create | cli.py:1126 | skill_create | Deprecated alias of crewai create skill: warns, then scaffolds an agent skill in ./skills/ or the current directory. |
crewai skill install | cli.py:1139 | skill_install | Downloads a registry skill named @org/name and installs it into ./skills/ inside a project, or under ~/.crewai/skills/ outside one. |
crewai skill publish | cli.py:1155 | skill_publish | Publishes the skill in the current directory, scoped to the user's organisation or --org; --force skips the git-state checks. |
crewai skill list | cli.py:1164 | skill_list | Lists the skills installed in the local cache. |
crewai skill search | cli.py:1174 | skill_search | Lists the organisation's skills on AMP, optionally filtered by a query. |
crewai template list | cli.py:1188 | template_list | Lists the available project templates and lets the user pick one to install. |
crewai template add | cli.py:1203 | template_add | Deprecated alias of crewai create template: warns, then copies a named template into the current directory. |
crewai flow kickoff | cli.py:1218 | flow_run | Deprecated: prints a warning and then runs the project exactly as crewai run does with no options. |
crewai flow plot | cli.py:1228 | flow_plot | Draws the project's flow as the interactive HTML page, from the declarative definition when there is one and otherwise through the project's uv plot script. |
crewai flow add-crew | cli.py:1238 | flow_add_crew | Backfills the project id and scaffolds a new crew inside an existing flow project. |
crewai triggers list | cli.py:1257 | triggers_list | Lists the triggers offered by the integrations connected to the user's organisation on AMP. |
crewai triggers run | cli.py:1267 | triggers_run | Fetches the sample payload of trigger app_slug/trigger_slug from AMP and runs the local crew with it. |
crewai org list | cli.py:1294 | org_list | Lists the organisations the logged-in user belongs to. |
crewai org switch | cli.py:1302 | switch | Makes the given organisation the active one in the CLI settings. |
crewai org current | cli.py:1309 | current | Shows which organisation is currently active. |
crewai enterprise configure | cli.py:1322 | enterprise_configure | Backfills the project id, fetches the OAuth2 settings published by an enterprise CrewAI AMP URL and writes them into the CLI settings. |
crewai config list | cli.py:1341 | config_list | Lists every CLI setting and its current value. |
crewai config set | cli.py:1352 | config_set | Sets one CLI setting to the given value. |
crewai config reset | cli.py:1361 | config_reset | Returns every CLI setting to its default. |
crewai env view | cli.py:1375 | env_view | Prints a table of the tracing variables (CREWAI_TRACING_ENABLED, CREWAI_TESTING, CREWAI_USER_ID, CREWAI_ORG_ID) and whether a .env file exists here. |
crewai traces enable | cli.py:1451 | traces_enable | Backfills the project id and records trace consent in the user data file, so later crew and flow runs send traces to the platform. |
crewai traces disable | cli.py:1477 | traces_disable | Records that the user declined tracing; CREWAI_TRACING_ENABLED=true in the environment still overrides it. |
crewai traces status | cli.py:1500 | traces_status | Shows the environment variable, the recorded consent and whether tracing is in effect. |
crewai checkpoint (bare) | cli.py:1548 | checkpoint | Resolves the checkpoint location (default ./.checkpoints, a directory or a SQLite file) for its subcommands, and with no subcommand opens the checkpoint terminal UI. |
crewai checkpoint list | cli.py:1564 | checkpoint_list | Lists the checkpoints in a directory or SQLite file. |
crewai checkpoint info | cli.py:1578 | checkpoint_info | Shows the details of one checkpoint, or of the latest when given a directory. |
crewai checkpoint resume | cli.py:1593 | checkpoint_resume | Resumes a run from the given checkpoint id, defaulting to the most recent. |
crewai checkpoint diff | cli.py:1605 | checkpoint_diff | Compares two checkpoints side by side. |
crewai checkpoint prune | cli.py:1626 | checkpoint_prune | Deletes old checkpoints, keeping the newest N or removing those older than a duration, with a dry-run option. |
crewai-devtools declares 5 console scripts: bump-version, tag, release, docs-check, devtools (lib/devtools/pyproject.toml lines 22 to 27). The package is private: both publishing workflows skip it. The devtools group also carries an agents-md-eval subcommand reached only through the devtools script, so it is not a console script of its own. Separately, eight Python files run as programs: seven under scripts/ and the tool spec generator in crewai-tools, which CI runs.
| Entry point | Declared at | Handler | What it does |
|---|---|---|---|
bump-version | lib/devtools/src/crewai_devtools/cli.py:2107 | bump | Sets a new version in every package under lib/, then commits and pushes unless told not to. It does not tag, publish or release. |
tag | lib/devtools/src/crewai_devtools/cli.py:2223 | tag | After the bump PR merges, reads the version from the packages and creates and pushes the git tag and a GitHub release with editable notes. |
release | lib/devtools/src/crewai_devtools/cli.py:2348 | release | Runs the whole release: a LiteLLM catalog refresh PR, the version bump PR (waiting for its merge), release notes, docs, tag and GitHub release, then the enterprise repository's release. Each phase can be skipped or resumed by flag. |
docs-check | lib/devtools/src/crewai_devtools/docs_check.py:352 | docs_check | Diffs the branch against a base ref and asks an OpenAI model which documentation the change needs. With --write it generates the English pages and their ar, ko and pt-BR translations. |
devtools | lib/devtools/src/crewai_devtools/cli.py:2618 | main | Runs the devtools command group, which holds bump, tag, release, docs-check and agents-md-eval as subcommands. |
scripts/age90_file_input_runner.py | scripts/age90_file_input_runner.py:201 | main | Manual runner for PDF input handling: builds multimodal payloads from a PDF in fallback, payload or full kickoff mode so the result can be inspected. |
scripts/docs/freeze_current_edge.py | scripts/docs/freeze_current_edge.py:42 | main | Freezes the current edge docs as the snapshot for a given version and registers it in docs.json; re-running for the same version does nothing. |
scripts/docs/freeze_historical_versions.py | scripts/docs/freeze_historical_versions.py:160 | main | Extracts the docs as they were at each listed release tag from git and writes them under docs/v<tag>/, skipping tags already frozen unless --force. |
scripts/docs/prefix_version_paths.py | scripts/docs/prefix_version_paths.py:396 | main | One-time migration of docs.json to directory-based versions: points each version at its snapshot, adds the Edge entry and writes redirects for the canonical URLs. |
scripts/jev_evaluation_runner.py | scripts/jev_evaluation_runner.py:573 | main | Scores a recorded dataset of routing and guardrail decisions offline against confidence thresholds and prints the report as JSON. |
scripts/update_litellm_catalog.py | scripts/update_litellm_catalog.py:26 | main | Refreshes the checked-in LiteLLM model catalog by hand; release runs the same update in its own PR. |
scripts/wharf_runner.py | scripts/wharf_runner.py:230 | main | Exercises SDK tracing end to end: runs one or more flows of three standalone agents that search and scrape through Firecrawl's hosted MCP server, printing each run's execution and trace ids. |
generate_tool_specs.py | lib/crewai-tools/src/crewai_tools/generate_tool_specs.py:194 | ToolSpecExtractor.extract_all_tools | Walks every exported tool class and builds its name, description, argument schema and required environment variables; the script then writes the result to tool.specs.json. |
These seventeen public methods are where an application hands work to CrewAI. Each family offers a synchronous call, a thread-wrapped async call and a native async call; on Flow and Agent the native one is an alias. All three classes accept a checkpoint to resume from. Flow adds Flow.resume and Flow.resume_async, which continue a run paused for human feedback, possibly in a later process (section 7 draws that lifecycle).
| Method | Declared at | What it does |
|---|---|---|
Crew.train | lib/crewai/src/crewai/crew.py:954 | Runs a copy of the crew n times collecting human feedback, then has an evaluator turn each agent's feedback into suggestions saved to the named pickle. On failure it clears both training files and re-raises. |
Crew.kickoff | lib/crewai/src/crewai/crew.py:1006 | Executes the crew's tasks with the given inputs and files and returns a CrewOutput, or a streaming output when streaming is on. It restores from a checkpoint first when one is given. |
Crew.kickoff_for_each | lib/crewai/src/crewai/crew.py:1110 | Runs a fresh copy of the crew once per input dictionary, one after another, and returns the outputs with usage summed across runs. |
Crew.kickoff_async | lib/crewai/src/crewai/crew.py:1146 | Async wrapper that runs the synchronous kickoff on a worker thread, with an async stream when streaming is on. |
Crew.kickoff_for_each_async | lib/crewai/src/crewai/crew.py:1200 | Runs the crew once per input concurrently through the thread-wrapped async kickoff and gathers the results. |
Crew.akickoff | lib/crewai/src/crewai/crew.py:1226 | Native async kickoff: task execution, memory and knowledge queries all use async and await instead of a worker thread. |
Crew.akickoff_for_each | lib/crewai/src/crewai/crew.py:1329 | Runs the native async kickoff once per input concurrently and gathers the results, or one merged stream when streaming. |
Crew.replay | lib/crewai/src/crewai/crew.py:2118 | Loads the stored outputs of the last kickoff and re-executes the crew from the given task onward, reusing the earlier outputs as context. It raises when the task id is not among the stored outputs. |
Crew.test | lib/crewai/src/crewai/crew.py:2353 | Runs a copy of the crew n times while an evaluator LLM scores every task, then prints the score table. |
Flow.resume | lib/crewai/src/crewai/flow/runtime/__init__.py:1349 | Synchronous entry that continues a flow paused for human feedback by running the async resume in a new event loop. It refuses to run inside a running loop. |
Flow.resume_async | lib/crewai/src/crewai/flow/runtime/__init__.py:1402 | Applies the human's feedback to the pending step (collapsing it to an outcome with an LLM when outcomes are declared), stores it and fires the downstream listeners. It returns the final output, or a new pending marker at the next feedback point. |
Flow.kickoff | lib/crewai/src/crewai/flow/runtime/__init__.py:2118 | Synchronous entry that runs the async kickoff in a new event loop, or on a one-thread pool when a loop is already running. It also handles checkpoint restore and streaming. |
Flow.kickoff_async | lib/crewai/src/crewai/flow/runtime/__init__.py:2183 | Restores persisted state when an id or source state is given, applies the inputs, emits the flow started event, runs every start method and its listeners, and returns the last method's output. |
Flow.akickoff | lib/crewai/src/crewai/flow/runtime/__init__.py:2771 | Alias that calls the async kickoff with the same arguments. |
Agent.kickoff | lib/crewai/src/crewai/agent/core.py:1952 | Runs one agent on its own, without a crew, on a message or message list with optional tools, structured output and files. Inside a running flow it returns a coroutine for the flow to await. |
Agent.kickoff_async | lib/crewai/src/crewai/agent/core.py:2344 | Native async form of the standalone run: prepares the executor, emits the started event, runs the reasoning loop and builds the agent output. |
Agent.akickoff | lib/crewai/src/crewai/agent/core.py:2429 | Alias that calls the async standalone run with the same arguments. |
The repository has 19 GitHub Actions workflows. They are entry points in the sense that a GitHub event starts work: pull requests start the checks, a push to main refreshes caches, cron starts the nightly build and housekeeping, and a maintainer starts a release by hand through publish.yml. CI runs ruff check, ruff format --check, mypy (strict, pydantic plugin) over lib/, pytest with --block-network --timeout=60 under xdist and pytest-randomly, pip-audit and CodeQL. The rows here say what starts each one; section 10 places the workflows in the deployment topology and is where their part in testing and release is described.
| Workflow | Trigger at | Trigger | What it does |
|---|---|---|---|
agents-md-eval.yml | .github/workflows/agents-md-eval.yml:9 | pull_request_target to main touching the template AGENTS.md or CLAUDE.md | Asks four coding-agent models what the scaffolded AGENTS.md teaches, with the base and the PR versions, and fails on a drop of more than 10%. It runs base-branch code only and skips forks, because it spends API keys. |
agents-md.yml | .github/workflows/agents-md.yml:7 | pull_request to main touching templates, the eval or linked docs | Runs the AGENTS.md guard tests and the eval's own tests, which the main test workflow skips. |
build-uv-cache.yml | .github/workflows/build-uv-cache.yml:3 | push to main changing uv.lock or pyproject.toml; cron every 5 days; manual | Installs every dependency group for Python 3.10 to 3.13 and saves the uv cache the other workflows restore. |
close-associated-prs.yml | .github/workflows/close-associated-prs.yml:3 | issues closed | When someone with triage rights or higher closes an issue as completed, closes the open pull requests in this repository linked to it. |
codeql.yml | .github/workflows/codeql.yml:14 | push and pull_request to main, ignoring templates | Runs CodeQL static analysis. |
docs-broken-links.yml | .github/workflows/docs-broken-links.yml:3 | pull_request or push to main touching docs/; manual | Prunes all but the edge and default doc versions and runs the Mintlify broken-link check. |
first-contributor-welcome.yml | .github/workflows/first-contributor-welcome.yml:3 | pull_request_target closed | Posts a thank-you comment with share links on a first-time contributor's merged pull request, once. |
ftc-require-issue.yml | .github/workflows/ftc-require-issue.yml:3 | pull_request_target opened, edited or reopened | Closes a first-time contributor's pull request that names no open issue, with a comment and a needs-issue label; authors with write access are exempt. |
generate-tool-specs.yml | .github/workflows/generate-tool-specs.yml:3 | pull_request to main touching crewai_tools; manual | Runs generate_tool_specs.py and commits a changed tool.specs.json back to the branch with a GitHub App token. |
linter.yml | .github/workflows/linter.yml:3 | pull_request | Runs ruff check and ruff format --check over lib/ when code changed. |
nightly.yml | .github/workflows/nightly.yml:3 | cron daily 06:00 UTC; manual | When main has commits from the last 25 hours, stamps a .devYYYYMMDD version on every package, builds them and publishes all but crewai-devtools to PyPI. |
pr-size.yml | .github/workflows/pr-size.yml:3 | pull_request opened, synchronized or reopened | Labels the pull request by the size of its diff. |
pr-title.yml | .github/workflows/pr-title.yml:3 | pull_request opened, edited, synchronized or reopened | Checks that the pull request title is a semantic (conventional commit) title. The step runs with continue-on-error, so a bad title is reported but does not fail the workflow. |
publish.yml | .github/workflows/publish.yml:3 | workflow_dispatch with an optional release_tag | Builds every package at the tag, publishes all but crewai-devtools to PyPI with the PyPI token, and posts the release notes to Slack. |
stale.yml | .github/workflows/stale.yml:8 | cron daily 12:10 UTC; manual | Marks issues stale after 30 days and closes them 5 days later; marks pull requests stale after 45 days without closing them. |
tests.yml | .github/workflows/tests.yml:3 | pull_request | Runs the crewai, tools, and CLI, core and files suites in eight duration-balanced groups per Python version with the network blocked and a 60 second timeout. |
type-checker.yml | .github/workflows/type-checker.yml:3 | pull_request | Runs mypy over lib/ for each Python version when code changed. |
update-test-durations.yml | .github/workflows/update-test-durations.yml:3 | push to main touching tests/**/*.py; manual | Runs the whole suite per Python version and caches the timings that tests.yml uses to balance its groups. |
vulnerability-scan.yml | .github/workflows/vulnerability-scan.yml:3 | pull_request; push to main; cron Mondays 09:00 UTC | Runs pip-audit over the installed dependencies and publishes the report as an artifact and job summary. |
crewai flow plot writes an HTML page that loads lib/crewai/src/crewai/flow/visualization/assets/interactive.js. Its event listeners are the entry points of that page: a click or key press in the browser starts work there, with no server behind it. The file has 19 matching lines; the one at line 2325 (an export-json button) is commented out, which leaves 18 live listeners. Each row names the method that registers the listener; the "What it does" column describes the listener that row adds. Two of them (the PNG and PDF exports) fetch scripts from cdnjs.cloudflare.com at click time, which section 12 lists.
| Event and element | Declared at | Registered by | What it does |
|---|---|---|---|
| resize on window | interactive.js:649 | setupCanvas | Resizes the highlight canvas to the new window size. |
| click on drawer overlay | interactive.js:1400 | setupEventListeners | Closes the node details drawer. |
| click on drawer close button | interactive.js:1401 | setupEventListeners | Closes the node details drawer. |
| keydown on document | interactive.js:1403 | setupEventListeners | Closes the drawer when the key is Escape. |
| click on a code copy button | interactive.js:1804 | attachContentEventListeners | Copies the method's source to the clipboard and marks the button as copied for two seconds. |
| click on an accordion header | interactive.js:1819 | attachContentEventListeners | Expands or collapses that section of the drawer. |
| click on a triggered-by link | interactive.js:1829 | attachContentEventListeners | Highlights the node that triggers this method. |
| click on a trigger group | interactive.js:1841 | attachContentEventListeners | Highlights every node in the group of triggers listed on the element. |
| click on a condition group | interactive.js:1854 | attachContentEventListeners | Highlights the trigger nodes of an AND or OR condition. |
| click on a router event link | interactive.js:1866 | attachContentEventListeners | Highlights the path a router takes for that event. |
| click on the router events title | interactive.js:1878 | attachContentEventListeners | Highlights every path out of the router. |
| change on the IDE selector | interactive.js:2251 | setupControls | Saves the preferred IDE for open-in-IDE links to local storage. |
| click on zoom in | interactive.js:2255 | setupControls | Zooms the graph in by 20% with a short animation. |
| click on zoom out | interactive.js:2263 | setupControls | Zooms the graph out by 20% with a short animation. |
| click on fit | interactive.js:2271 | setupControls | Fits the whole graph into the space left beside the legend and drawer. |
| click on export PNG | interactive.js:2279 | setupExportControls | Loads html2canvas 1.4.1 from cdnjs, renders the graph and downloads it as flow.png. |
| click on export PDF | interactive.js:2296 | setupExportControls | Loads html2canvas 1.4.1 and jspdf 2.5.1 from cdnjs, renders the graph and downloads it as flow.pdf. |
| click on the theme toggle | interactive.js:2460 | setupTheme | Switches between light and dark themes, saves the choice to local storage and recolours the edges. |
The evidence's entry-point search returned 114 matches across 37 files. Two groups of them are in the inventory above: the 18 live listeners, and the 8 script entry points (each of the seven files under scripts/ matched twice, once for its main function and once for its __main__ block). The remaining matches are not entry points of this system: cli.py line 1635 and the devtools cli.py lines 2618 and 2623 are already counted as the CLI and devtools rows; the @task decorators and the __main__ block under crewai_cli/templates/ belong to projects that crewai create generates, not to this repository; triggers/main.py lines 82 to 98 are dictionary lookups on a variable called app (app.get calls); and the rest are recorded HTTP cassettes, tool READMEs, docstring examples in the ChromaDB client, items in the devtools AGENTS.md eval data, and test files (including the commented-out export-json listener above). The library entry points, the console scripts and the CLI commands were taken from the source and the pyproject files rather than from that search, which does not look for them.
A filesystem has no modules, so the boundaries here are drawn by path prefix from the call and import graph, using the grouping rule modules.cjs prints on its second line, verbatim: grouped by path prefix under ./, starting at depth 2; a group over 25 files is split one level deeper, and only children with 8+ files are promoted. It ran with its defaults (--big 25, --min-child 8), which put 1543 files into 35 module groups. The variable depth is why lib/crewai/src/crewai (552 files) sits as one row beside test directories of 8 to 89 files: the source packages have no subdirectory of their own large enough to be promoted, while the test trees split by area.
The repository is a uv workspace of six packages, all at version 1.15.41: crewai, crewai-core, crewai-tools, crewai-files, crewai-cli and crewai-devtools. Five of them appear below as a source group under lib/<package>/src, a test group (or several) under lib/<package>/tests, and a one-file group holding the package's pyproject.toml. crewai-devtools is small enough (14 files) to stay whole as lib/devtools. The library proper is lib/crewai/src/crewai, and the packages split out of it are lib/crewai-core/src/crewai_core (shared platform plumbing), lib/cli/src/crewai_cli (the command line), lib/crewai-tools/src/crewai_tools (the tool catalogue) and lib/crewai-files/src/crewai_files (file inputs and provider uploads).
One row per group, in the order modules.cjs prints them (by file count), with what each owns. To see the files and symbols inside a group, with their callers and callees, open the symbol index and filter by path.
| Files | Module | Kind | What it owns |
|---|---|---|---|
| 552 | lib/crewai/src/crewai | Source package | The framework itself: Agent, Task, Crew, Flow and the AgentExecutor, LLM providers, memory, knowledge, RAG, events and tracing, telemetry, A2A, MCP, hooks, skills, checkpoint state, plus the deprecated crewai.cli shim and re-exports of crewai_core. |
| 268 | lib/crewai-tools/src/crewai_tools | Source package | The tool catalogue (84 tool directories), MCP and enterprise adapters, the RAG adapter, AWS Bedrock toolkits and the SSRF-validating HTTP security layer. |
| 97 | lib/cli/src/crewai_cli | Source package | The crewai command line: project scaffolding and templates, run, train, test, replay, deploy, tools and skills publishing, triggers, org and config commands, authentication and the Textual TUIs. |
| 89 | lib/crewai/tests | Tests | Top-level framework tests: crews, flows, checkpoints, streaming, human feedback, multimodal input, SQLite lifecycle and project loading. |
| 55 | lib/cli/tests | Tests | Tests for crewai_cli commands: run, deploy, tools, templates, settings and the TUIs. |
| 52 | lib/crewai-tools/tests/tools | Tests | One test module per tool in the crewai_tools catalogue. |
| 43 | lib/crewai/tests/llms | Tests | LLM provider tests (OpenAI, Anthropic, Azure, Bedrock, Gemini and the rest), mostly against recorded cassettes. |
| 41 | lib/crewai/tests/utilities | Tests | Tests for the shared utilities: agent helpers, converters, the event bus wiring, file handling and prompts. |
| 36 | lib/crewai-files/src/crewai_files | Source package | File inputs for agents: resolution, processing and validation, spreadsheet parsing, provider uploaders (Anthropic, Bedrock, Gemini, OpenAI) and the upload cache. |
| 30 | lib/crewai-core/src/crewai_core | Source package | Plumbing shared by the library and the CLI: OAuth2 device-flow auth and providers, the token manager, CLI settings, storage paths, the PlusAPI platform client, telemetry helpers and platform app catalogues. |
| 27 | lib/crewai/tests/telemetry | Tests | Tests for anonymous telemetry spans and their opt-out switches. |
| 25 | lib/crewai/tests/agents | Tests | Agent and AgentExecutor tests: the reasoning loop, tool use, retries and guardrails. |
| 23 | lib/crewai/tests/cli | Tests | CLI tests that reach crewai_core through crewai's re-export modules (and, in 8 files, import crewai_cli directly), copying lib/cli/tests (section 17). None imports the deprecated crewai.cli shim. |
| 22 | lib/crewai-tools/tests/rag | Tests | Tests for the crewai_tools RAG layer: loaders, chunkers and the adapter. |
| 20 | lib/crewai/tests/rag | Tests | Tests for the library's RAG clients and embedding configuration. |
| 15 | lib/crewai/tests/memory | Tests | Tests for unified memory, the LanceDB store and kickoff task output storage. |
| 15 | lib/crewai-files/tests | Tests | Tests for crewai_files processing, uploaders and the upload cache. |
| 14 | lib/devtools | Maintainer tooling | The crewai-devtools package and its tests: the bump-version, tag, release, docs-check and devtools console scripts, and the LiteLLM catalogue updater. |
| 13 | lib/crewai/tests/tools | Tests | Tests for BaseTool, tool usage and the tool result cache. |
| 13 | lib/crewai/tests/mcp | Tests | Tests for the MCP client, transports and tool resolution. |
| 12 | lib/crewai/tests/skills | Tests | Tests for agent skills loading and use. |
| 12 | lib/crewai/tests/hooks | Tests | Tests for the before and after LLM-call and tool-call hooks. |
| 12 | lib/crewai-core/tests | Tests | Tests for crewai_core settings, auth and the platform client. |
| 11 | lib/crewai/tests/experimental | Tests | Tests for the experimental modules (evaluation and the agent-level experiments). |
| 10 | lib/crewai-tools/tests | Tests | Top-level crewai_tools tests: base tool, file and URL readers, safe path and safe request helpers, the MCP adapter and tool spec generation. |
| 9 | lib/crewai/tests/events | Tests | Tests for the event bus and event types. |
| 9 | lib/cli/tests/authentication | Tests | Tests for CLI login: the OAuth2 device flow, providers and token storage. |
| 8 | lib/crewai/tests/tracing | Tests | Tests for execution trace batches sent to the CrewAI platform. |
| 4 | scripts | Scripts | Standalone runners and maintenance scripts: evaluation runners, a file-input runner and the LiteLLM catalogue update. |
| 1 | (repo root) | Test configuration | The workspace conftest.py shared by every test suite. |
| 1 | lib/crewai-files | Package manifest | The crewai-files pyproject.toml. |
| 1 | lib/crewai | Package manifest | The crewai pyproject.toml, which also declares the crewai console script. |
| 1 | lib/cli | Package manifest | The crewai-cli pyproject.toml, which declares the same crewai console script. |
| 1 | lib/crewai-core | Package manifest | The crewai-core pyproject.toml. |
| 1 | lib/crewai-tools | Package manifest | The crewai-tools pyproject.toml. |
modules.cjs prints 20 cross-module edges of 89 (69 below the top 20). It also ignored 2235 edges with an end outside ./, which it describes as mostly test-to-source and a measure of coverage rather than coupling. Every one of the 20 printed edges runs from a test group to a source group, so test-to-source dominates edge weight: the heaviest edge, lib/crewai/tests into lib/crewai/src/crewai, carries 2049 calls and imports. Coupling between the source packages sits in the 69 edges below the cut, and the two-way pairs below are where it shows. The individual calls behind any edge can be traced in the symbol index.
| Count | From | To |
|---|---|---|
| 2049 | lib/crewai/tests | lib/crewai/src/crewai |
| 664 | lib/crewai/tests/agents | lib/crewai/src/crewai |
| 631 | lib/cli/tests | lib/cli/src/crewai_cli |
| 499 | lib/crewai/tests/utilities | lib/crewai/src/crewai |
| 498 | lib/crewai-files/tests | lib/crewai-files/src/crewai_files |
| 382 | lib/crewai/tests/llms | lib/crewai/src/crewai |
| 334 | lib/crewai/tests/hooks | lib/crewai/src/crewai |
| 274 | lib/crewai/tests/telemetry | lib/crewai/src/crewai |
| 239 | lib/crewai/tests/events | lib/crewai/src/crewai |
| 225 | lib/crewai-tools/tests/rag | lib/crewai-tools/src/crewai_tools |
| 213 | lib/crewai-tools/tests/tools | lib/crewai-tools/src/crewai_tools |
| 191 | lib/crewai/tests/skills | lib/crewai/src/crewai |
| 178 | lib/crewai/tests/tools | lib/crewai/src/crewai |
| 173 | lib/crewai-tools/tests | lib/crewai-tools/src/crewai_tools |
| 160 | lib/crewai/tests/rag | lib/crewai/src/crewai |
| 147 | lib/crewai/tests/cli | lib/crewai-core/src/crewai_core |
| 120 | lib/crewai/tests/memory | lib/crewai/src/crewai |
| 98 | lib/cli/tests/authentication | lib/crewai-core/src/crewai_core |
| 77 | lib/crewai/tests/tracing | lib/crewai/src/crewai |
| 73 | lib/crewai/tests/telemetry | lib/crewai-core/src/crewai_core |
modules.cjs, so it matches the tables in this section exactly. Columns are dependency depth: a module sits one column right of everything that calls into it. Line thickness is the number of calls and imports crossing that boundary. The 10 modules under “no edge drawn” take part in none of the edges shown, heaviest first. The accented box is the module with the most files.The figure draws 34 of the 35 groups. The one it leaves out is (repo root), the single conftest.py.
A layered diagram cannot show a cycle, so the cycles are listed here. modules.cjs prints 11 two-way pairs, meaning the dependency runs both ways. Seven involve a test group. In those, the reverse count is small (1 to 9). The graph produced those edges by matching calls by name: a call to a generic method name on a passed-in callback is attributed to a same-named function in a test file. No source module imports a test module. Four of the two-way pairs are between source packages: crewai <-> crewai_core (60 / 3), crewai_cli <-> crewai (53 / 2), crewai_tools <-> crewai (37 / 4), crewai <-> crewai_files (25 / 12). These four matter, because the packages were split out of crewai so that each would depend on the core and not the other way round. The source shows only two of them are real: crewai_core and crewai_files never import crewai (a search for from crewai. in both returns nothing), and their reverse edges are the graph resolving the stdlib import logging to the module crewai/a2a/utils/logging.py.
| One side | Other side | Forward / reverse | What the source shows |
|---|---|---|---|
| lib/crewai/tests | lib/crewai/src/crewai | 2049 / 3 | Test to source. The 3 reverse edges are calls inside the LLM call wrappers, the retry helper and the project wrappers, resolved by name to functions defined in test files. |
| lib/cli/tests | lib/cli/src/crewai_cli | 631 / 1 | Test to source, with one reverse name match. |
| lib/crewai/tests/llms | lib/crewai/src/crewai | 382 / 2 | Test to source, with two reverse name matches. |
| lib/crewai-tools/tests/tools | lib/crewai-tools/src/crewai_tools | 213 / 1 | Test to source, with one reverse name match. |
| lib/crewai/tests | lib/crewai/tests/agents | 63 / 2 | Two test trees. The edges between them are calls resolved by name to same-named helpers in the other tree. |
| lib/crewai/src/crewai | lib/crewai-core/src/crewai_core | 60 / 3 | Real one way only. crewai imports crewai_core (auth, settings, paths, PlusAPI, telemetry helpers). crewai_core contains no import of crewai. Its reverse edges are its stdlib import logging resolved by the graph to crewai/a2a/utils/logging.py, plus method names matched by name. |
| lib/cli/src/crewai_cli | lib/crewai/src/crewai | 53 / 2 | A real cycle. The CLI imports the framework to run crews and flows. The framework imports the CLI back in one place: flow/conversational_mixin.py:685-686 lazily imports the interactive prompt and the conversational flow TUI from crewai_cli. |
| lib/crewai-tools/src/crewai_tools | lib/crewai/src/crewai | 37 / 4 | A real cycle. Tools subclass the framework's BaseTool and import its RAG and MCP types. The framework calls back into crewai_tools to build platform tools for an agent and for JSON crews, to wrap MCP servers in project/crew_base.py, and to fetch spreadsheet inputs through the SSRF-safe bounded GET. |
| lib/crewai/src/crewai | lib/crewai-files/src/crewai_files | 25 / 12 | Real one way only. crewai imports crewai_files for file inputs and uploads. crewai_files contains no import of crewai. Its 12 reverse edges come from its own stdlib import logging lines, which the graph resolves to crewai/a2a/utils/logging.py. |
| lib/crewai/tests | lib/cli/tests | 23 / 9 | Calls resolved by name across the two suites, for example a CLI test's agent factory matched to a helper of the same name in a crew thread-safety test. Not shared fixtures. (The overlap between lib/crewai/tests/cli and lib/cli/tests is a separate matter, covered in section 17.) |
| lib/crewai/tests/agents | lib/crewai/tests/tools | 2 / 1 | One tool-failure test call resolved by name to a recording LLM class in an agent test, and the reverse. |
So the two genuine cycles between packages are crewai with crewai_cli and crewai with crewai_tools. In both, the dependency back into crewai is a lazy import inside a function, which avoids an import-time cycle but still means neither package can be installed and used without the other.
This is the backbone of a run, taken from graphdata.json and checked against the source. A crew run goes down six layers: CLI, Crew, Task, Agent, Executor, LLM. The Flow runtime is drawn beside them because it is not a separate path. AgentExecutor is itself a Flow (class AgentExecutor(Flow[AgentExecutorState], BaseAgentExecutor)), and its reasoning loop is a set of @start, @listen and @router methods, so every agent turn runs on the same machinery as a user's flow.
Solid edges are in the graph. Dashed edges are calls that are in the source but that the graph does not resolve, because they are calls on another object rather than on self. Each dashed edge was found by reading the source:
_execute_tasks calls execute_sync on each synchronous task (crew.py:1643)._execute_core calls execute_task on the task's agent (task.py:894)._execute_without_timeout calls invoke on the agent executor (agent/core.py:1143).invoke_async awaits the executor's own Flow kickoff_async (agent_executor.py:3288)._call_llm_retrying_empty runs a closure that get_llm_response built around call on the LLM (agent_utils.py:766).The step from run_crew in run_crew.py down to kickoff on Crew is not an edge at all. crewai run, install and related commands spawn uv run subprocesses inside the user's project, and the project's own entry point calls kickoff there. That process boundary is drawn in section 6.
Two more calls in the source loop back up the layers without new functions. When a guardrail rejects an output, _invoke_guardrail_function calls execute_task on the agent again (task.py:1508). And _execute_with_timeout runs _execute_without_timeout on a worker thread, so a timed task reaches invoke by the same route as an untimed one. A synchronous invoke outside an event loop calls the executor's own Flow kickoff, and inside a loop it hands back the invoke_async coroutine, so both entry points end in the Flow column.
Every function in Figure 2, one row each, in the order the figure draws them.
| Function | Owner | Where | Layer | What it does |
|---|---|---|---|---|
run | crewai_cli | lib/cli/src/crewai_cli/cli.py:754 | CLI | The click handler for crewai run. It rejects a trained-agents file combined with a flow definition, then passes the file, definition and inputs to run_crew. |
run_crew | crewai_cli | lib/cli/src/crewai_cli/cli.py:54 | CLI | A thin wrapper that imports the real run_crew from the crewai_cli run_crew module on first use and forwards its arguments, keeping the CLI module cheap to load. |
run_crew | crewai_cli | lib/cli/src/crewai_cli/run_crew.py:1137 | CLI | Works out what the project is (declarative flow, JSON crew, Python flow, agent or classic crew), asks for any missing model API key, and starts it inside the project's environment. The run itself happens in a uv run subprocess. |
kickoff | Crew | lib/crewai/src/crewai/crew.py:1006 | Crew | Runs the crew: resumes from a checkpoint when asked, prepares inputs and agents, runs the tasks under the sequential or hierarchical process, and returns a CrewOutput (or a streaming output). A failure emits a kickoff-failed event. |
prepare_kickoff | crews/utils.py | lib/crewai/src/crewai/crews/utils.py:243 | Crew | Gets a crew ready to run: before-kickoff callbacks, the start event, input interpolation into tasks and agents, task callbacks, agent setup and planning. |
_run_sequential_process | Crew | lib/crewai/src/crewai/crew.py:1533 | Crew | Runs the crew's tasks in their declared order by handing them all to _execute_tasks. |
_run_hierarchical_process | Crew | lib/crewai/src/crewai/crew.py:1537 | Crew | Creates the manager agent that delegates the work (or validates the one supplied), then runs the tasks through _execute_tasks. |
_execute_tasks | Crew | lib/crewai/src/crewai/crew.py:1586 | Crew | Walks the task list from the start or the replay index, runs synchronous tasks with execute_sync and collects asynchronous ones as futures, and builds the CrewOutput. Each finished synchronous task goes to _process_task_result and its execution log is stored for replay. |
_process_task_result | Crew | lib/crewai/src/crewai/crew.py:1906 | Crew | Writes a completed task's output to the crew's output log file when one is configured. |
execute_sync | Task | lib/crewai/src/crewai/task.py:626 | Task | Records the start time and runs the task on the calling thread through _execute_core. |
_execute_core | Task | lib/crewai/src/crewai/task.py:850 | Task | Runs the task with its agent: emits the start event, calls the agent's execute_task, builds the TaskOutput, applies the guardrails and saves any output file. A failure emits TaskFailedEvent before the error propagates. |
_invoke_guardrail_function | Task | lib/crewai/src/crewai/task.py:1425 | Task | Checks a task output against one guardrail and, when it fails, asks the agent to redo the task, up to guardrail_max_retries more attempts. It raises when the attempts run out. |
execute_task | Agent | lib/crewai/src/crewai/agent/core.py:1003 | Agent | The public entry for an agent working one task. It opens a scope for any spreadsheet inputs and delegates to _execute_task. |
_execute_task | Agent | lib/crewai/src/crewai/agent/core.py:1031 | Agent | Builds the task prompt (knowledge retrieval, spreadsheets, tools), emits the agent-started event, and runs the executor with or without a timeout depending on max_execution_time. Errors other than a timeout go to the retry handler, bounded by max_retry_limit. |
_execute_with_timeout | Agent | lib/crewai/src/crewai/agent/core.py:1090 | Agent | Runs _execute_without_timeout on a worker thread and raises TimeoutError if it has not finished within max_execution_time seconds. |
_execute_without_timeout | Agent | lib/crewai/src/crewai/agent/core.py:1130 | Agent | Calls the agent executor's invoke with the prompt and tool descriptions and returns its output. It refuses to run synchronously inside a running event loop. |
invoke | AgentExecutor | lib/crewai/src/crewai/agents/agent_executor.py:3132 | Executor | The synchronous entry to the reasoning loop. Inside an event loop it returns the invoke_async coroutine; otherwise it resets the executor state and runs the loop by calling the Flow kickoff on itself. |
invoke_async | AgentExecutor | lib/crewai/src/crewai/agents/agent_executor.py:3229 | Executor | The asynchronous entry to the reasoning loop. It sets up messages and file inputs, awaits the executor's own Flow kickoff_async, and returns the final answer. |
call_llm_and_parse | AgentExecutor | lib/crewai/src/crewai/agents/agent_executor.py:1643 | Executor | The @router step that runs on continue_reasoning: it calls the model through get_llm_response, parses the answer into an action or a final answer, and routes to parsed, parser_error or context_error. |
get_llm_response | agent_utils | lib/crewai/src/crewai/utilities/agent_utils.py:713 | LLM | Prepares the messages (before-call hooks, and summarising the context when it is near the window limit), calls the model through _call_llm_retrying_empty, and runs the after-call hooks. |
_call_llm_retrying_empty | agent_utils | lib/crewai/src/crewai/utilities/agent_utils.py:529 | LLM | Makes the LLM call and, if the answer is empty, repeats it once after admitting it through the RPM limiter. It raises EmptyLLMResponseError if the retry is empty too. |
enforce_rpm_limit | agent_utils | lib/crewai/src/crewai/utilities/agent_utils.py:463 | LLM | Calls the requests-per-minute limiter when one is configured. The limiter sleeps until the next minute once the limit is reached. |
call | BaseLLM | lib/crewai/src/crewai/llms/base_llm.py:541 | LLM | The abstract method every LLM provider implements: it sends the messages, and any tool schemas, to the model and returns text, tool calls or a structured response. |
kickoff | Flow | lib/crewai/src/crewai/flow/runtime/__init__.py:2118 | Flow runtime | The synchronous entry for a flow. It runs kickoff_async through _run_flow under asyncio.run, on a one-worker thread pool when an event loop is already running. |
_run_flow | Flow | lib/crewai/src/crewai/flow/runtime/__init__.py:2164 | Flow runtime | A coroutine defined inside kickoff that awaits kickoff_async with the inputs, files and restore id. |
kickoff_async | Flow | lib/crewai/src/crewai/flow/runtime/__init__.py:2183 | Flow runtime | Restores state from persistence when an id is given, applies the inputs, emits the flow-started event and runs every start method. It emits the flow-finished event and returns the final output. |
_execute_start_method | Flow | lib/crewai/src/crewai/flow/runtime/__init__.py:2837 | Flow runtime | Runs one @start method (or, when resuming, skips one that already completed) and then triggers the listeners waiting on it. |
_execute_method | Flow | lib/crewai/src/crewai/flow/runtime/__init__.py:2927 | Flow runtime | Runs one flow method between started and finished (or failed) events. When the method raises HumanFeedbackPending, it saves the paused state through persistence, defaulting to default_flow_persistence. |
default_flow_persistence | flow/persistence/factory.py | lib/crewai/src/crewai/flow/persistence/factory.py:48 | Flow runtime | Returns the backend from a registered persistence factory, or a built-in SQLiteFlowPersistence when none is registered. |
_execute_listeners | Flow | lib/crewai/src/crewai/flow/runtime/__init__.py:3171 | Flow runtime | After a method completes, runs the triggered routers one at a time and then the triggered listeners in parallel. When a router's result triggers a start method, it re-enters _execute_start_method, which is how cyclic flows and the agent loop repeat. |
Most of CrewAI runs inside one Python process, and inside that process the call graph in section 5 is the whole story. Nine flows leave it: they cross into another operating-system process, onto the network, onto another thread, or across time into a later process that picks up where an earlier one stopped. The table names each one and the boundary it crosses. Four of them are drawn as sequence diagrams, because the hop is something the call graph cannot follow: the crewai run subprocess, A2A delegation, human feedback pause and resume, and telemetry and trace export. The other five are described in their row only, since each is one request and one answer, or one thread and one join.
| Flow | Boundary crossed | What crosses it | Where |
|---|---|---|---|
| crewai run subprocess | Process: the CLI spawns uv run inside the user's project | The project script name, and an environment built from os.environ plus private-index credentials. Only the exit code comes back. Drawn in Figure 6.1. | lib/cli/src/crewai_cli/run_crew.py:1348-1383 |
| LLM provider call | Network: HTTPS to the provider's API | Messages, tool schemas and the API key. Every provider's call and acall is wrapped by run_with_rate_limit_retry: 3 attempts, 1 s initial delay, 8 s cap, 20% jitter, only for 429/throttling errors, and a provider's Retry-After hint wins over the computed delay. | lib/crewai/src/crewai/llms/retry.py:13-37,124; base_llm.py:246-284 |
| OAuth2 device login | Network, plus a human in a browser | _get_device_code asks the identity provider for a device code; the user approves in the browser; _poll_for_token polls the token URL at the server's interval until the code's lifetime ends (600 s if the provider gives none); validate_jwt_token checks the JWT against the provider's JWKS and save_tokens writes it encrypted to disk. | lib/crewai-core/src/crewai_core/auth/oauth2.py:111-200 |
| Deploy to CrewAI AMP | Network: HTTPS to https://app.crewai.com (or CREWAI_PLUS_URL) | A bearer token and either the git remote URL (_create_payload) or a zip of the project (_create_crew_from_zip), plus the values in the project's .env. Build and run happen on the platform, outside this repository; deploy redeploys and the status and log commands read back. | lib/cli/src/crewai_cli/deploy/main.py:432,567,722,793 |
| Telemetry and trace export | Thread and network: spans leave from a background exporter thread; trace batches go to the platform | Anonymous OpenTelemetry spans to https://telemetry.crewai.com:4319, and, when tracing is on, a buffered batch of execution events sent through PlusAPI at the end of the run. Drawn in Figure 6.4. | lib/crewai-core/src/crewai_core/telemetry.py:307-315; trace_batch_manager.py:91-386 |
| A2A delegation | Network: HTTP JSON-RPC (default), gRPC or HTTP+JSON to a remote agent; a thread when called from a running event loop | The agent card fetch, a user message, then streamed, polled or pushed task updates until a terminal or input-needing state. Drawn in Figure 6.2. | lib/crewai/src/crewai/a2a/utils/delegation.py:135-650 |
| MCP tool call | Process for stdio servers (a spawned subprocess); network for HTTP and SSE servers | Tool discovery and tool invocations. _retry_operation gives tool discovery and each tool call up to 3 attempts with 1 s then 2 s waits, each attempt under a 30 s timeout, and does not retry connection or authentication errors; the connect step itself is not retried (section 13). A new client and transport is opened per tool call. | lib/crewai/src/crewai/mcp/client.py:53-56,702-754; mcp/transports/ |
| Human feedback pause and resume | Time across processes: state is written to SQLite in one process and read back by another | Flow state, method outputs and the pending feedback context, keyed by the flow id. Drawn in Figure 6.3. | lib/crewai/src/crewai/flow/runtime/__init__.py:1260-1693,2452-2506; flow/persistence/sqlite.py:285-383 |
| Async task execution | Thread: each async_execution task runs on its own daemon thread | The task, its context (only the last synchronous output) and a copy of the context variables go out; a Future carries the result back. execute_async starts the thread and _process_async_tasks joins the futures in order before the next synchronous task. kickoff_async runs the whole synchronous kickoff on a worker thread, while akickoff is native asyncio throughout. | lib/crewai/src/crewai/task.py:650-679; crew.py:1146-1198,1226,1586-1655,2035 |
The crewai command a developer types is not the process that runs their crew. run hands to run_crew, which reads the project's pyproject.toml, works out whether it is a crew, a Python flow, a declarative flow or a JSON crew, and for a classic crew calls _execute_uv_script. That function builds the child environment with build_env_with_all_tool_credentials (the parent's environment plus UV_INDEX_* credentials for each private package index the project names) and runs uv run run_crew. uv resolves the project's own virtual environment and starts the project's script there, so the crew imports the project's pinned dependencies, not the CLI's. JSON crews take the same path through _project_run_command, with the crew definition and inputs passed as environment variables. Nothing but the exit code comes back: on a non-zero code _handle_run_error prints what it can and the CLI exits with the child's code.
An agent configured with remote A2A agents delegates through _delegate_to_a2a, which runs up to the configured maximum number of turns and calls execute_a2a_delegation once per turn. That synchronous wrapper drives the async implementation with asyncio.run, on a single worker thread when an event loop is already running in the caller, so delegation crosses a thread boundary as well as the network. _aexecute_a2a_delegation_impl fetches the remote agent card through _afetch_agent_card_cached (an HTTP GET of /.well-known/agent-card.json, cached for 300 s and retried on 401 by retry_on_401), checks the configured client auth against the card with validate_auth_against_agent_card, negotiates a transport (JSON-RPC unless the card and config agree on gRPC or HTTP+JSON), attaches auth headers and sends the message. One of three update handlers then waits for the task: streaming by default, polling through _poll_task_until_complete (2 s interval, 300 s timeout by default), or push notifications. Whatever arrives is reduced by process_task_state; an input-required answer goes back to the local agent, which may send another turn.
A Flow method decorated with human_feedback asks a feedback provider for a human's answer. A synchronous provider blocks; an asynchronous one raises HumanFeedbackPending, which the runtime treats as a pause rather than an error. _execute_method emits a paused event instead of a failed one, and the kickoff path writes the flow state and the pending context to SQLite through save_pending_feedback, then returns the pending object to the caller and lets the process end. Hours later, possibly on another machine sharing the storage directory, a new process calls from_pending with the flow id, which reads the row back with load_pending_feedback and rebuilds the flow. resume (or resume_async) then runs _resume_async_body, which applies the feedback, marks the method complete, deletes the pending row with clear_pending_feedback, and continues with the method's listeners. If a later listener pauses again, the cycle repeats with a new pending row.
Two separate streams leave a run. Anonymous telemetry is on by default: the Telemetry class records spans such as crew_creation (versions, process type, counts, agent roles, tool names; goals, backstories, task text and inputs only when the crew sets share_crew), and an OpenTelemetry batch processor exports them from a background thread over OTLP HTTP to https://telemetry.crewai.com:4319 with a 30 s timeout. It is off when OTEL_SDK_DISABLED, CREWAI_DISABLE_TELEMETRY or CREWAI_DISABLE_TRACKING is set. Execution traces are the second stream, decided by should_enable_tracing (an explicit tracing argument, then CREWAI_TRACING_ENABLED, then the user's stored consent). When on, initialize_batch opens a batch on the CrewAI platform through initialize_trace_batch (or its ephemeral variant for a user who is not logged in), every event on the bus is buffered by add_event, and at crew completion, failure or a termination signal finalize_batch sends the whole buffer with send_trace_events in one request and closes the batch with finalize_trace_batch.
| Function | Where | What it does |
|---|---|---|
run | lib/cli/src/crewai_cli/cli.py:754 | The click handler for crewai run. It parses the options and calls run_crew. |
run_crew | lib/cli/src/crewai_cli/run_crew.py:1137 | Reads the project's pyproject.toml, decides which kind of project it is, prompts for missing model keys, and dispatches to the right runner. |
_execute_uv_script | lib/cli/src/crewai_cli/run_crew.py:1348 | Runs a project script through uv run as a subprocess with the augmented environment. A non-zero exit becomes ProjectScriptFailed with the child's code. |
build_env_with_all_tool_credentials | lib/crewai-core/src/crewai_core/tool_credentials.py:35 | Copies os.environ and adds UV_INDEX credentials for each private package index named under tool.uv.sources. Errors reading the project file are swallowed and the plain copy is returned. |
_project_run_command | lib/cli/src/crewai_cli/run_crew.py:596 | Builds the command that runs inline Python inside the project's environment: uv run --no-sync python -c, or poetry run for a Poetry project. |
_handle_run_error | lib/cli/src/crewai_cli/run_crew.py:1386 | Prints the subprocess error and any captured output. It adds a hint to run crewai update when the project still uses Poetry. |
call | lib/crewai/src/crewai/llms/base_llm.py:541 | The synchronous LLM call each provider class implements. BaseLLM wraps every subclass's version with the rate-limit retry when the subclass is defined. |
acall | lib/crewai/src/crewai/llms/base_llm.py:611 | The asynchronous counterpart of call, wrapped with the async rate-limit retry in the same way. |
run_with_rate_limit_retry | lib/crewai/src/crewai/llms/retry.py:124 | Runs an LLM call up to three times, sleeping between attempts only when the error is a provider throttle. Nested calls inside an active retry run once, so retries do not multiply. |
_get_device_code | lib/crewai-core/src/crewai_core/auth/oauth2.py:111 | Posts the client id, scopes and audience to the provider's device authorisation URL and returns the device code, user code and verification URL. |
_poll_for_token | lib/crewai-core/src/crewai_core/auth/oauth2.py:136 | Polls the token URL at the server's interval until the user approves or the code expires. It continues on authorization_pending and slow_down and exits with an error on timeout. |
validate_jwt_token | lib/crewai-core/src/crewai_core/auth/utils.py:13 | Verifies the access token's RS256 signature against the provider's JWKS and checks issuer, audience, expiry and the required claims. |
save_tokens | lib/crewai-core/src/crewai_core/token_manager.py:50 | Encrypts the access token and its expiry with Fernet and writes it atomically to tokens.enc with owner-only permissions. |
deploy | lib/cli/src/crewai_cli/deploy/main.py:432 | Handles crewai deploy push: validates the project, reads its deployment status, and redeploys from git or uploads a fresh zip depending on how AMP holds the code. |
create_crew | lib/cli/src/crewai_cli/deploy/main.py:567 | Handles crewai deploy create: registers a new deployment on AMP from the git remote or, without one, from a zip. |
_create_payload | lib/cli/src/crewai_cli/deploy/main.py:793 | Builds the create request body: deployment name, the git clone URL and the variables from the project's .env. |
_create_crew_from_zip | lib/cli/src/crewai_cli/deploy/main.py:722 | Archives the project and posts it as a multipart upload to AMP, with the .env values as form fields. |
crew_creation | lib/crewai/src/crewai/telemetry/telemetry.py:254 | Records the telemetry span for a new crew: versions, process type, counts and agent and tool names, plus prompts and inputs only when share_crew is set. |
should_enable_tracing | lib/crewai/src/crewai/events/listeners/tracing/utils.py:144 | Decides whether execution traces are collected, from an explicit override, then CREWAI_TRACING_ENABLED, then eval runs, then the user's stored consent. |
initialize_batch | lib/crewai/src/crewai/events/listeners/tracing/trace_batch_manager.py:91 | Creates the local trace batch and opens it on the platform, authenticated or ephemeral. On a 401 or 403 it falls back to ephemeral. |
add_event | lib/crewai/src/crewai/events/listeners/tracing/trace_batch_manager.py:283 | Appends one trace event to the in-memory buffer. Nothing is sent at this point. |
finalize_batch | lib/crewai/src/crewai/events/listeners/tracing/trace_batch_manager.py:327 | At the end of a run, waits for pending handlers, sends all buffered events and closes the batch. A timeout marks the batch as failed. |
initialize_trace_batch | lib/crewai-core/src/crewai_core/plus_api.py:599 | POSTs a new trace batch to the CrewAI platform's tracing API with the user's bearer token. |
send_trace_events | lib/crewai-core/src/crewai_core/plus_api.py:616 | POSTs a list of trace events to an open batch on the platform. |
finalize_trace_batch | lib/crewai-core/src/crewai_core/plus_api.py:636 | PATCHes a batch on the platform to mark it finished. |
_delegate_to_a2a | lib/crewai/src/crewai/a2a/wrapper.py:1233 | Runs the multi-turn delegation loop for an agent that has remote A2A agents, calling the delegation once per turn up to the configured maximum. |
execute_a2a_delegation | lib/crewai/src/crewai/a2a/utils/delegation.py:135 | Synchronous wrapper that runs the async delegation with asyncio.run. If an event loop is already running, it does this on a one-thread executor. |
_aexecute_a2a_delegation_impl | lib/crewai/src/crewai/a2a/utils/delegation.py:375 | One delegation turn: fetches the card, checks auth and extensions, negotiates the transport, builds the message, and runs the chosen update handler. |
_afetch_agent_card_cached | lib/crewai/src/crewai/a2a/utils/agent_card.py:227 | Fetches and validates a remote agent card over HTTP, caching it for 300 seconds. |
retry_on_401 | lib/crewai/src/crewai/a2a/auth/utils.py:194 | Retries a request that got a 401, waiting with exponential backoff and reapplying the client auth scheme before each new attempt. |
validate_auth_against_agent_card | lib/crewai/src/crewai/a2a/auth/utils.py:149 | Checks that the client auth configured for a remote agent matches a security scheme its card declares. |
_poll_task_until_complete | lib/crewai/src/crewai/a2a/updates/polling/handler.py:45 | Calls get_task on the remote agent at a fixed interval until the task is terminal or needs input. It raises a timeout error when the time or poll limit is reached. |
process_task_state | lib/crewai/src/crewai/a2a/task_helpers.py:138 | Turns a remote task's state and messages into the result the local agent sees: completed with text, input or auth required, failed (rejected folds in here), canceled, or nothing yet. |
_retry_operation | lib/crewai/src/crewai/mcp/client.py:702 | Runs an MCP operation up to three times under a per-attempt timeout, with exponential waits. Connection, authentication and not-found errors are not retried. |
execute_async | lib/crewai/src/crewai/task.py:650 | Starts the task on a daemon thread with a copy of the current context variables and returns a Future for its output. |
_process_async_tasks | lib/crewai/src/crewai/crew.py:2035 | Waits on each pending async task's Future in order, then records the output and the execution log as for a synchronous task. |
kickoff_async | lib/crewai/src/crewai/crew.py:1146 | Runs the synchronous crew kickoff on a worker thread with asyncio.to_thread, so async callers can await it. |
akickoff | lib/crewai/src/crewai/crew.py:1226 | Native async kickoff: tasks, memory and knowledge calls are awaited directly instead of running the sync path on a thread. |
human_feedback | lib/crewai/src/crewai/flow/human_feedback.py:453 | Decorator that marks a Flow method as needing a human's answer on its output before the flow continues. |
_execute_method | lib/crewai/src/crewai/flow/runtime/__init__.py:2927 | Runs one Flow method with its events and hooks. A pending-feedback signal is reported as a pause rather than a failure and re-raised. |
save_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:285 | Saves the flow state row and inserts or replaces the pending_feedback row for the flow id. |
from_pending | lib/crewai/src/crewai/flow/runtime/__init__.py:1260 | Class method that rebuilds a paused flow from persistence by flow id: it restores state, method outputs and the pending context, and raises if nothing is pending. |
load_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:329 | Reads the pending_feedback row for a flow id and returns the saved state and context, or None. |
resume | lib/crewai/src/crewai/flow/runtime/__init__.py:1349 | Synchronous resume with the human's feedback. It runs resume_async under asyncio.run and refuses if an event loop is already running. |
resume_async | lib/crewai/src/crewai/flow/runtime/__init__.py:1402 | Async resume. It requires a pending context, runs the resume body, and emits a flow-failed event if anything raises. |
_resume_async_body | lib/crewai/src/crewai/flow/runtime/__init__.py:1469 | Applies the feedback, marks the paused method complete, deletes the pending row, and runs the downstream listeners to completion or the next pause. |
clear_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:368 | Deletes the pending_feedback row for a flow id. |
Three genuine state machines exist in the source, each with named states and explicit transitions: the AgentExecutor reasoning loop, the Flow human feedback lifecycle, and the A2A task lifecycle. Each is drawn below with a table of its states. Elsewhere the code has status fields (a task's output being present or not, a trace batch marked failed), but none has a set of states and transitions a reader could name, so none is drawn.
AgentExecutor is itself a Flow (class AgentExecutor(Flow[AgentExecutorState], BaseAgentExecutor), at lib/crewai/src/crewai/agents/agent_executor.py:203), so the agent's reasoning loop is written as Flow methods decorated with @start, @listen and @router. Each router returns a string, and the Flow runtime's _execute_listeners fires whichever methods listen for that string or for the method's own name. The returned strings are the machine's states. After the run starts with generate_plan and initialize_reasoning, each iteration goes the same way: check_max_iterations either forces a final answer (max_iter, 25 by default, reached, or the same tool called too many times in a row) or sends the agent to one of two reasoning paths. Text tool calling goes through call_llm_and_parse and route_by_answer_type. Native function calling goes through call_llm_native_tools and execute_native_tool. Tool results, parser errors and context-length errors all return to initialized for the next iteration, and the loop ends in agent_finished, which finalize turns into the agent's output.
Twelve core states are drawn. A second set handles planning when it is enabled: check_todos_available returns planning_disabled, no_todos or has_todos, and the todo machinery adds single_todo_ready, multiple_todos_ready, todo_injected, step_executed, step_observed_low, step_observed_medium, step_observed_high, continue_plan, replan_now, refine_and_continue, goal_achieved, needs_replan, todo_satisfied, todo_not_satisfied, todo_marked and all_todos_complete. Without planning, the tool paths pass through check_todo_completion or check_native_todo_completion, which return todo_not_satisfied, and increment_and_continue returns initialized. Two more strings end the loop beside agent_finished: native_finished (a native-tools answer with no tool calls) and tool_result_is_final (a tool flagged as returning the final answer).
| State | Returned by | What it means | Leaves to |
|---|---|---|---|
| initialized | initialize_reasoning, increment_and_continue, recover_from_parser_error, recover_from_context_length | A new iteration is about to start. | continue_iteration, then check_max_iterations: continue_reasoning, continue_reasoning_native or force_final_answer |
| continue_reasoning | check_max_iterations; also call_llm_native_tools when the model turns out not to support native tools | Ask the LLM for a ReAct-style text answer and parse it. | call_llm_and_parse: parsed, parser_error or context_error |
| continue_reasoning_native | check_max_iterations when the agent uses native function calling | Ask the LLM with tool schemas attached. | call_llm_native_tools: native_tool_calls, native_finished, context_error or continue_reasoning |
| parsed | call_llm_and_parse | The text answer was parsed into a tool action or a final answer. | route_by_answer_type: execute_tool or agent_finished |
| execute_tool | route_by_answer_type | The parsed action names a tool to run. | execute_tool_action: tool_completed or tool_result_is_final |
| native_tool_calls | call_llm_native_tools | The model returned one or more structured tool calls. | execute_native_tool: native_tool_completed or tool_result_is_final |
| tool_completed | execute_tool_action | The tool's result has been added to the conversation. | check_todo_completion, then increment_and_continue: initialized |
| native_tool_completed | execute_native_tool | The native tool results have been added to the conversation. | check_native_todo_completion, then increment_and_continue: initialized |
| parser_error | call_llm_and_parse | The model's text could not be parsed as an action or a final answer. | recover_from_parser_error: initialized, with an error message added for the model |
| context_error | call_llm_and_parse, call_llm_native_tools | The request exceeded the model's context window. | recover_from_context_length: summarises the history (or raises when respect_context_window is off), then initialized |
| force_final_answer | check_max_iterations | The iteration cap or the repeated-tool-call limit was reached. | ensure_force_final_answer: agent_finished |
| agent_finished | route_by_answer_type, ensure_force_final_answer | A final answer is available. | finalize: completed (terminal) |
A Flow with a human_feedback step can stop in the middle and continue later in a different process (the sequence is Figure 6.3). The states are the flow's, seen from outside: running, paused, resumed, completed and failed. Paused is the HumanFeedbackPending signal, persisted as a row in the pending_feedback table of the flow SQLite store (section 11). The source has no event or field named resumed: a resume is visible as a fresh flow-started event followed by a method-started event for the method that paused, inside _resume_async_body.
| State | Entered when | Visible as | What is persisted |
|---|---|---|---|
| running | kickoff, kickoff_async or akickoff starts the flow | FlowStartedEvent, then a started and finished event per method | The flow_states row after each method, when persistence is configured |
| paused | An asynchronous feedback provider raises HumanFeedbackPending inside a human_feedback step | MethodExecutionPausedEvent, then FlowPausedEvent; kickoff returns the pending object and the flow's pending_feedback property is set | save_pending_feedback writes the state row and the pending_feedback row (context, state, created_at), keyed by flow_uuid |
| resumed | A later process calls from_pending with the flow id, then resume or resume_async with the feedback | No dedicated event: FlowStartedEvent again and a method-started event for the paused method | clear_pending_feedback deletes the pending row once the feedback is applied; the state row stays |
| completed | The last listener finishes, in the first process or after a resume | FlowFinishedEvent | Final state row, when persistence is configured |
| failed | A method raises anything other than HumanFeedbackPending, before or after a resume | MethodExecutionFailedEvent and FlowFailedEvent, emitted by _emit_flow_failed | Nothing new; if the failure comes during a resume before the delete, the pending row is kept and the resume can be retried |
A task on a remote A2A agent moves through the protocol's eight task states. The state values come from the a2a SDK. crewai groups them in lib/crewai/src/crewai/a2a/task_helpers.py:38-58: submitted and working are pending, input_required and auth_required are actionable (the remote side wants something), and completed, failed, canceled and rejected are terminal. The update handlers wait until a task is terminal or actionable, and process_task_state maps that state to what the delegating agent sees (Figure 6.2). When crewai is the server, the code it runs itself sets only two states: completed, at the end of _execute_impl, and canceled, in cancel. A failure raises a server error and leaves the failed state to the SDK, and submitted and working are also set by the SDK.
| State | Group | Who sets it | How crewai's client handles it |
|---|---|---|---|
| submitted | Pending | The remote server, through the a2a SDK, when it accepts the message | Keeps waiting: streaming skips the update, polling polls again |
| working | Pending | The remote server, through the a2a SDK | Keeps waiting until the polling or push timeout |
| input_required | Actionable | The remote agent, when it needs more from the caller | Returned with the remote agent's message; the delegation loop may send another turn |
| auth_required | Actionable | The remote agent, when the caller must authenticate further | Returned as auth_required to the delegating agent |
| completed | Terminal | The remote server; crewai's own server code sets it at the end of _execute_impl | Returned as completed with the result text and artifacts |
| failed | Terminal | The remote server; crewai's server raises a server error and leaves the state to the SDK; crewai's handlers also build failed results on transport errors and timeouts | Returned as failed with the error |
| canceled | Terminal | The remote server; crewai's own server code sets it in cancel | Returned as canceled |
| rejected | Terminal | The remote server, when it refuses the task | Folded into failed by process_task_state |
| Function | Where | What it does |
|---|---|---|
_execute_listeners | lib/crewai/src/crewai/flow/runtime/__init__.py:3171 | After a Flow method finishes, it starts every method that listens for that method's name or, for a router, for the string it returned. |
generate_plan | lib/crewai/src/crewai/agents/agent_executor.py:573 | The executor's start method. It builds a plan of todos when planning is enabled and does nothing otherwise. |
check_todos_available | lib/crewai/src/crewai/agents/agent_executor.py:1281 | Routes after planning to planning_disabled, no_todos or has_todos. |
initialize_reasoning | lib/crewai/src/crewai/agents/agent_executor.py:1604 | Prepares the reasoning loop and returns initialized. |
continue_iteration | lib/crewai/src/crewai/agents/agent_executor.py:2514 | A bridge listener on initialized that hands each new iteration to the iteration check. |
check_max_iterations | lib/crewai/src/crewai/agents/agent_executor.py:2520 | Returns force_final_answer when max_iter or the repeated-tool-call limit is reached. Otherwise it picks the native or text reasoning path. |
call_llm_and_parse | lib/crewai/src/crewai/agents/agent_executor.py:1643 | Calls the LLM for a text answer and parses it into a tool action or a final answer. It returns parsed, parser_error or context_error. |
route_by_answer_type | lib/crewai/src/crewai/agents/agent_executor.py:1906 | Sends a parsed tool action to execute_tool and a final answer to agent_finished, or to the todo path when planning is on. |
call_llm_native_tools | lib/crewai/src/crewai/agents/agent_executor.py:1724 | Calls the LLM with tool schemas and routes on whether it returned tool calls, a final answer or a context error. It falls back to text reasoning when the model does not support native tools. |
execute_tool_action | lib/crewai/src/crewai/agents/agent_executor.py:1920 | Runs the tool named by a parsed action and appends its result. It returns tool_result_is_final when the tool's result is the answer. |
execute_native_tool | lib/crewai/src/crewai/agents/agent_executor.py:2010 | Runs the structured tool calls the model returned and appends their results to the conversation. |
check_todo_completion | lib/crewai/src/crewai/agents/agent_executor.py:2535 | After a text-path tool run, decides whether the current todo is satisfied. With no todo it returns todo_not_satisfied. |
check_native_todo_completion | lib/crewai/src/crewai/agents/agent_executor.py:2498 | The native-path counterpart: any tool run satisfies an active todo, and with no todo it returns todo_not_satisfied. |
increment_and_continue | lib/crewai/src/crewai/agents/agent_executor.py:2646 | Counts the iteration and returns initialized to start the next one. |
recover_from_parser_error | lib/crewai/src/crewai/agents/agent_executor.py:3094 | Adds a parser-error message for the model, counts the iteration and returns initialized. |
recover_from_context_length | lib/crewai/src/crewai/agents/agent_executor.py:3117 | Summarises the conversation to fit the context window (or raises when that is disabled), counts the iteration and returns initialized. |
ensure_force_final_answer | lib/crewai/src/crewai/agents/agent_executor.py:1619 | Asks the model for a final answer from what it has so far and returns agent_finished. |
finalize | lib/crewai/src/crewai/agents/agent_executor.py:2663 | Listens for any finishing state and turns the current answer into the executor's result. It returns completed, or skipped if the run already finished. |
human_feedback | lib/crewai/src/crewai/flow/human_feedback.py:453 | Decorator that marks a Flow method as needing a human's answer on its output before the flow continues. |
_resume_async_body | lib/crewai/src/crewai/flow/runtime/__init__.py:1469 | Applies the feedback, marks the paused method complete, deletes the pending row, and runs the downstream listeners to completion or the next pause. |
kickoff | lib/crewai/src/crewai/flow/runtime/__init__.py:2118 | Starts a flow synchronously, running its async runtime under asyncio.run. It returns the final output or, on a pause, the pending object. |
save_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:285 | Saves the flow state row and inserts or replaces the pending_feedback row for the flow id. |
from_pending | lib/crewai/src/crewai/flow/runtime/__init__.py:1260 | Class method that rebuilds a paused flow from persistence by flow id. It raises if nothing is pending. |
resume | lib/crewai/src/crewai/flow/runtime/__init__.py:1349 | Synchronous resume with the human's feedback. It runs the async resume under asyncio.run. |
resume_async | lib/crewai/src/crewai/flow/runtime/__init__.py:1402 | Async resume. It requires a pending context and emits a flow-failed event if the resume raises. |
clear_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:368 | Deletes the pending_feedback row for a flow id. |
_emit_flow_failed | lib/crewai/src/crewai/flow/runtime/__init__.py:2730 | Emits FlowFailedEvent with the error so listeners and traces record the failure. |
process_task_state | lib/crewai/src/crewai/a2a/task_helpers.py:138 | Turns a remote task's state and messages into the delegating agent's result: completed, input or auth required, failed (including rejected), canceled, or nothing yet. |
_execute_impl | lib/crewai/src/crewai/a2a/utils/task.py:306 | Server side: runs the local agent on an incoming A2A task and enqueues the task as completed with its artifacts. On an exception it emits a failure event and raises a server error. |
cancel | lib/crewai/src/crewai/a2a/utils/task.py:458 | Server side: marks a running task canceled and enqueues the final status update. It does nothing if the task is already terminal. |
CrewAI is a library and a CLI. It runs inside the developer's own process with that process's rights, serves nothing on the network by default, and holds no users of its own. Its security model is therefore about what it reaches out to and what it lets an agent do: the credentials it reads, the hosts it calls, the code and data it loads from disk, and the guards it puts between a model's output and the machine. The CrewAI AMP platform, which holds accounts and deployments, is outside this repository; the CLI is its client.
Six boundaries matter. Everything inside the developer's process is trusted equally: an agent, its tools and the developer's own code share one address space, one set of environment variables and one filesystem view. Everything a model returns is untrusted input to that process, which is why the guards in the controls table sit where a model's output becomes a URL, a SQL statement or a callable.
| Boundary | What crosses it | What protects it |
|---|---|---|
| The developer's process | Environment variables (every provider and tool key), the working directory, local stores (section 11) | Operating-system user permissions only. The library adds no isolation between agents, tools and the host code. |
| LLM providers | Prompts, task text, tool results and memory snippets out; completions and tool calls back | HTTPS and the provider's API key. Model output is treated as data, then parsed into actions the guards below constrain. |
| Tool APIs and URLs an agent may fetch | Queries and URLs chosen by the model; page and API content back into the prompt | The SSRF guard on URL-fetching tools and loaders; each vendor tool's own API key; nothing filters returned content. |
| MCP servers | Tool calls and results; for stdio servers, a subprocess the developer configured | Whatever the server's transport offers. A stdio server runs with the developer's rights. |
| Remote A2A agents | Delegated messages out; task states, text and artifacts back | Client auth schemes (bearer, basic, digest, API key, OAuth2 client credentials and authorisation code, mutual TLS) and a check that they match the remote agent card. |
| The CrewAI platform | Login, deployments (git URL or zip, plus .env values), trace batches, anonymous telemetry spans | OAuth2 device-flow login and a bearer token; the platform enforces its own authorisation and answers 401 or 403. |
The CLI. crewai login uses the OAuth2 device flow (RFC 8628). The identity source is a configurable provider, workos by default, with auth0, entra_id, keycloak and okta also supported, and the platform URL defaults to https://app.crewai.com. _get_device_code obtains a device and user code, the user approves in a browser, and _poll_for_token waits for the token. validate_jwt_token then verifies the RS256 signature against the provider's JWKS and checks issuer, audience, expiry, not-before and issued-at with 10 s leeway. Token lifetime is the JWT's own exp claim: save_tokens records it, get_token treats the token as absent once it has passed, and no refresh token is stored, so an expired session means logging in again. get_auth_token is what every platform call uses, and PlusAPI sends the token as a bearer header with the organisation id.
Token storage. The token is Fernet-encrypted in tokens.enc, with the key generated by _get_or_create_key and kept in secret.key in the same directory (~/.local/share/crewai/credentials on Linux), which is created 0700, with both files written 0600. Because the key and the ciphertext sit side by side, the encryption stops a casual read of one file, not someone who can read the user's files. CLI settings, including the organisation, OAuth2 settings and the tool repository username and password, are written as JSON to ~/.config/crewai/settings.json with mode 0600.
Provider API keys. Model and tool keys come from environment variables (for example OPENAI_API_KEY, read by _normalize_openai_fields) or from constructor arguments, and are sent only to the provider they belong to. A missing key is found when the client is built: _get_client_params raises MissingAPIKeyError at that point, not at startup. crewai run passes the whole environment to the project subprocess (section 6).
A2A servers. For an agent exposed over A2A, crewai defines six server auth schemes in lib/crewai/src/crewai/a2a/auth/server_schemes.py: SimpleTokenAuth, EnterpriseTokenAuth, OIDCAuth, OAuth2ServerAuth, APIKeyServerAuth and MTLSServerAuth, each with an authenticate method. OIDCAuth and OAuth2ServerAuth validate JWTs against a JWKS (OAuth2ServerAuth can also use RFC 7662 token introspection). SimpleTokenAuth and APIKeyServerAuth compare a shared secret. MTLSServerAuth accepts every request and relies on the TLS layer having checked the client certificate. EnterpriseTokenAuth raises NotImplementedError. The schemes are configuration for whatever HTTP server hosts the agent: crewai builds and signs the agent card, but no code in its src tree runs an A2A HTTP server or calls authenticate: the only reference to a scheme is the A2AServerConfig.auth field (a2a/config.py:684). So five of the six schemes are implemented, none is invoked by crewai itself, and enforcement is left to the server the developer mounts the agent in.
There is none inside the library. No role, permission or access check exists in lib/crewai/src/crewai: an agent can call any tool it was given, every tool acts with whatever credentials the environment holds, and code an agent reaches runs with the developer's operating-system rights. Authorisation is delegated to two places. The CrewAI platform enforces its own, answering 401 or 403 to the CLI and to trace uploads (the trace manager then falls back to an anonymous ephemeral batch). And a small set of opt-in guards limit what model output can do locally: these are the controls in the table below. When a developer does not narrow an agent's tools, nothing else does, so what an agent can do is decided entirely by the tools and keys the developer hands it.
Nothing is served. What leaves by default is outbound only: anonymous telemetry spans to https://telemetry.crewai.com:4319 unless OTEL_SDK_DISABLED, CREWAI_DISABLE_TELEMETRY or CREWAI_DISABLE_TRACKING is set; the LiteLLM model price catalog, a JSON file the CLI fetches from raw.githubusercontent.com; and, when a developer opens the flow visualisation page, vis-network 9.1.2 from unpkg.com and html2canvas 1.4.1 and jspdf 2.5.1 from cdnjs.cloudflare.com, pinned by version but without subresource integrity. Execution traces leave only when tracing is enabled. The outbound evidence found no install hooks, and its three non-test risk-pattern lines are all that price catalog, which is data, not code. Section 12 gives each host and flagged line its row.
| Control | What it does | Default | Where |
|---|---|---|---|
| OAuth2 device flow | CLI login against a configurable identity provider. The JWT is verified against the provider's JWKS (RS256; issuer, audience and time claims) before it is stored. | On for every platform command; workos provider | lib/crewai-core/src/crewai_core/auth/oauth2.py:97-200; auth/utils.py:13 |
| Token storage | Fernet-encrypted tokens.enc with an owner-only key file beside it. Expiry is enforced on read; there is no refresh. | Always | lib/crewai-core/src/crewai_core/token_manager.py:19-136 |
| Provider API keys | Read from the environment or constructor arguments and sent only to the matching provider. A missing key raises when the client is first built. | Environment | lib/crewai/src/crewai/llms/providers/openai/completion.py:422-444,563-570 |
| A2A server authentication | Six scheme classes for an agent exposed over A2A: shared token, enterprise token (not implemented), OIDC, OAuth2 (JWKS or introspection), API key, mutual TLS. | Off unless configured; not invoked by crewai's own code | lib/crewai/src/crewai/a2a/auth/server_schemes.py:108-744 |
| SSRF guard | Resolves the host, refuses any non-global address (private ranges, loopback, link-local and the cloud metadata addresses), connects to the checked address, then checks the connected peer again. Each redirect hop is validated, credentials are stripped across origins, and proxies are refused. | On for URL-fetching tools and RAG loaders. CREWAI_TOOLS_ALLOW_UNSAFE_PATHS turns it off and CREWAI_TOOLS_ALLOWED_PRIVATE_HOSTS allowlists hosts. | lib/crewai-tools/src/crewai_tools/security/ssrf_adapter.py:60-316; safe_path.py:201; safe_requests.py:129 |
| NL2SQL DML guard | Read-only by default. It rejects multiple statements, comments, writable CTEs, EXPLAIN ANALYZE of a write and any write command, and rolls back non-write transactions. | Read-only; writes need allow_dml or CREWAI_NL2SQL_ALLOW_DML=true | lib/crewai-tools/src/crewai_tools/tools/nl2sql/nl2sql_tool.py:234-372 |
| Callback deserialisation | A callback stored as a dotted path (for example in a checkpoint) is only imported and resolved when the developer opts in. Otherwise the string is refused. | Off; needs CREWAI_DESERIALIZE_CALLBACKS | lib/crewai/src/crewai/types/callback.py:26-75 |
| Training pickles | crewai train writes training_data.pkl and trained_agents_data.pkl to the working directory, and agents load them with pickle on every run, with no integrity check. | Loaded whenever the file is present | lib/crewai/src/crewai/utilities/file_handler.py:120-170; agent/core.py:1529,1556 |
| Telemetry | Anonymous spans with versions, counts, agent roles and tool names. Goals, backstories, task text and inputs are added only when a crew sets share_crew. | On; three environment variables turn it off | lib/crewai-core/src/crewai_core/telemetry.py:45,286-359; lib/crewai/src/crewai/telemetry/telemetry.py:254 |
The SSRF guard covers less than all tool HTTP. The adapter does check the connected peer IP, but only the callers of security/safe_requests.py go through it: the website scraping, URL read and element scraping tools and the RAG web, PDF, DOCX and docs-site loaders (8 files in crewai-tools), plus crewai's own spreadsheet input fetch. Vendor API tools call their fixed hosts with requests or httpx directly.
Around the code, the repository adds supply-chain controls: CodeQL and pip-audit run in GitHub Actions, ruff runs the bandit (S) rules over lib/, and uv resolves with exclude-newer = "3 days" (a supply-chain cooldown) and pins named packages past known CVEs (pyproject.toml [tool.uv]).
| Function | Where | What it does |
|---|---|---|
_get_device_code | lib/crewai-core/src/crewai_core/auth/oauth2.py:111 | Posts the client id, scopes and audience to the provider's device authorisation URL. It returns the device code, user code and verification URL. |
_poll_for_token | lib/crewai-core/src/crewai_core/auth/oauth2.py:136 | Polls the token URL at the server's interval until the user approves or the device code expires (600 s when the provider states no lifetime). |
validate_jwt_token | lib/crewai-core/src/crewai_core/auth/utils.py:13 | Verifies the access token's RS256 signature against the provider's JWKS and checks issuer, audience, time claims and the required claims. |
save_tokens | lib/crewai-core/src/crewai_core/token_manager.py:50 | Encrypts the access token and its expiry with Fernet and writes it atomically to tokens.enc with owner-only permissions. |
get_token | lib/crewai-core/src/crewai_core/token_manager.py:65 | Decrypts tokens.enc and returns the access token, or None when there is none or it has expired. |
_get_or_create_key | lib/crewai-core/src/crewai_core/token_manager.py:32 | Reads the Fernet key from secret.key in the credentials directory, creating it exclusively with mode 0600 on first use. |
get_auth_token | lib/crewai-core/src/crewai_core/auth/token.py:12 | Returns the stored platform token, or raises an authentication error telling the user to log in. |
_normalize_openai_fields | lib/crewai/src/crewai/llms/providers/openai/completion.py:422 | Fills the OpenAI model's API key from OPENAI_API_KEY and its base URL from OPENAI_BASE_URL or OPENAI_API_BASE when they were not passed in. |
_get_client_params | lib/crewai/src/crewai/llms/providers/openai/completion.py:563 | Assembles the OpenAI client's parameters and raises MissingAPIKeyError if no key was found. |
authenticate | lib/crewai/src/crewai/a2a/auth/server_schemes.py:118 | The abstract method each A2A server scheme implements: it checks a presented credential and returns the authenticated user or raises a 401 (503 when a JWKS cannot be reached). |
crewai login | lib/cli/src/crewai_cli/cli.py:914 | The CLI command that clears stored user settings and runs the device-flow login, then lets the user pick an organisation. |
Good. The controls are deliberate for a library that runs with the caller's rights: an SSRF-validating HTTP path for URL-fetching tools, NL2SQL writes and callback deserialisation off unless opted in, OAuth2 device-flow login, CodeQL, pip-audit, ruff bandit rules and a three-day dependency cooldown with CVE pins, and every outbound host is one an LLM framework is expected to call. The gaps are known and local: the CLI token's key sits beside the token, telemetry is opt-out, the visualisation pulls pinned CDN scripts without SRI, and trained-agent data is loaded with pickle. The technical assessment gives the reasoning per criterion.
CrewAI has no configuration file of its own for the library. Every setting is either a constructor argument (an LLM(api_key=...), a tool's api_key field) or an environment variable read at the point of use, usually with the argument taking precedence and the variable as the fallback. Importing crewai also loads a .env file from the working directory (python-dotenv's loader runs at module level in lib/crewai/src/crewai/llm.py:90 and project/crew_base.py:130), and crewai run loads the project's .env into the project subprocess it spawns. The CLI keeps its own small settings file, ~/.config/crewai/settings.json, for the platform URL, OAuth2 provider and organisation, and its login token in an encrypted file beside it (see "Secrets" below).
The evidence counted 150 distinct environment variables read through an accessor or declared in a .env sample; CREWAI_DISABLE_TRACKING is read in lib/crewai-core/src/crewai_core/telemetry.py:358 but is not in that list, so the inventory below carries 151 rows, one per variable. The scope column separates what a user of the library sets (Library, LLM provider, Tool, Files, Core, CLI) from what only CI, the maintainers' devtools, a script or the test suite reads; none were dropped. Six names in the evidence are Python constants that hold a variable name rather than variables themselves (AWAITING_ENV, EVAL_TOKEN_ENV, PASTED_KEYS_ENV, RUN_COUNTED_ENV, TRACING_ENV_VAR, and CREWAI_TRAINED_AGENTS_FILE_ENV); their rows name the real variable they stand for. Where the evidence lists a variable that no code reads (it appears only in a README or a recorded cassette), the row says so.
The variables group into a few families. Telemetry and tracing: OTEL_SDK_DISABLED, CREWAI_DISABLE_TELEMETRY and CREWAI_DISABLE_TRACKING each switch off the anonymous OpenTelemetry export, which is on by default; CREWAI_TRACING_ENABLED opts a run in or out of sending execution traces to the CrewAI platform. Storage: CREWAI_STORAGE_DIR names the per-project directory under the user data directory where the SQLite, LanceDB and checkpoint stores live, defaulting to the name of the current directory. Model selection and provider keys: MODEL, OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY and the rest, read when a provider client is built. Tool keys: one or two per third-party tool in crewai-tools. Platform: CREWAI_PLUS_URL, the personal access and integration tokens. Opt-in guards: CREWAI_DESERIALIZE_CALLBACKS and CREWAI_NL2SQL_ALLOW_DML keep unsafe behaviour off unless set. CLI plumbing: variables the CLI sets for the project subprocess it starts and reads back there.
| Variable | Default | Read at | Scope | When absent |
|---|---|---|---|---|
A2A_TEST_ENDPOINT | http://localhost:9999 | lib/crewai/tests/a2a/test_a2a_integration.py:16 | Test-only | The A2A integration tests target a local agent on port 9999. |
ACTOR_LOGIN | none (set by the workflow) | .github/workflows/ftc-require-issue.yml:48 | CI-only | Set at line 34 from the event sender; if missing the inline script raises KeyError and the check step fails. |
AGENTOPS_API_KEY | none | lib/crewai/tests/cassettes/test_crew_kickoff_streaming_usage_metrics.yaml | Not read by code | Nothing: the name appears only inside recorded LLM response text in two cassettes. |
AGENT_HANDLER_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/merge_agent_handler_tool/merge_agent_handler_tool.py:65 | Tool | The Merge Agent Handler tool has no key unless one is passed; its calls to Merge cannot authenticate. |
ANTHROPIC_API_KEY | none | lib/crewai/src/crewai/llms/providers/anthropic/completion.py:457 | LLM provider | Building an Anthropic client without an explicit key raises MissingAPIKeyError. Also a CI secret in agents-md-eval.yml:53. |
APIFY_API_TOKEN | none | lib/crewai-tools/src/crewai_tools/tools/apify_actors_tool/apify_actors_tool.py:58 | Tool | The Apify Actors tool has no token unless one is passed and cannot call Apify. |
API_BASE | none | lib/crewai/src/crewai/utilities/llm_utils.py:532 | Library | The default LLM is built with its provider's own base URL; the CLI model picker also falls through to OLLAMA_HOST and then localhost:11434 (model_catalog.py:303). |
AUTHOR_ASSOCIATION | "" | .github/workflows/ftc-require-issue.yml:52 | CI-only | Treated as no association, so the author gets no maintainer exemption from the linked-issue rule. |
AUTHOR_LOGIN | none (set by the workflow) | .github/workflows/ftc-require-issue.yml:47 | CI-only | Set at line 33 from the PR author; if missing the script raises KeyError. |
AUTH_TOKEN | none | lib/crewai/src/crewai/a2a/auth/server_schemes.py:151 | Library (A2A server) | A SimpleTokenAuth scheme built without a token has no expected value to compare against. |
AWAITING_ENV | "" (stands for CREWAI_EVAL_AWAITING_RUN) | lib/crewai-core/src/crewai_core/eval_handshake.py:48,80 | Core (constant name) | The run is not one an evaluation is waiting on, so no evaluation handshake happens. |
AWS_ACCESS_KEY_ID | none | lib/crewai/src/crewai/llms/providers/bedrock/completion.py:300 | LLM provider | None is passed to boto3, which falls back to its own credential chain (profile, instance role); with none, Bedrock calls fail. |
AWS_DEFAULT_REGION | us-east-1 (Bedrock LLM); us-east-1 or us-west-2 in tools | lib/crewai/src/crewai/llms/providers/bedrock/completion.py:311 | LLM provider, Tool, Files | Bedrock falls back to AWS_REGION_NAME, then us-east-1; the knowledge-base retriever uses us-east-1 and the agent invoker us-west-2. |
AWS_REGION_NAME | us-east-1 | lib/crewai/src/crewai/llms/providers/bedrock/completion.py:312 | LLM provider | Region defaults to us-east-1. |
AWS_SECRET_ACCESS_KEY | none | lib/crewai/src/crewai/llms/providers/bedrock/completion.py:303 | LLM provider | As AWS_ACCESS_KEY_ID: boto3's credential chain decides. |
AZURE_API_BASE | none | lib/crewai/src/crewai/llms/providers/azure/completion.py:136 | LLM provider | Third fallback for the Azure endpoint after AZURE_ENDPOINT and AZURE_OPENAI_ENDPOINT; with none, the Azure client has no endpoint. |
AZURE_API_KEY | none | lib/crewai/src/crewai/llms/providers/azure/completion.py:131 | LLM provider | The Azure client uses azure-identity credentials, or for the OpenAI-shaped delegate falls back to OPENAI_API_KEY (line 211). |
AZURE_API_VERSION | 2024-06-01 | lib/crewai/src/crewai/llms/providers/azure/completion.py:139 | LLM provider | API version 2024-06-01 is used. |
AZURE_CREDENTIAL_SCOPES | none | lib/crewai/src/crewai/llms/providers/azure/completion.py:173 | LLM provider | The client's default token scopes are used. |
AZURE_DOCUMENTDB_CONNECTION_STRING | "" | lib/crewai-tools/src/crewai_tools/tools/azure_documentdb_vector_search_tool/vector_search.py:58 | Tool | No connection unless connection_string is passed to the tool. |
AZURE_DOCUMENTDB_E2E_CONNECTION_STRING | none | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:21 | Test-only | The local end-to-end tests are skipped. |
AZURE_DOCUMENTDB_LIVE_CONNECTION_STRING | none | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:151 | Test-only | The live Azure tests are skipped. |
AZURE_DOCUMENTDB_LIVE_EMBEDDING_DEPLOYMENT | none | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:159 | Test-only | Read only inside the live test, which is skipped without the connection string. |
AZURE_DOCUMENTDB_LIVE_IDENTITY | environment | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:168 | Test-only | The live test authenticates with the environment credential. |
AZURE_DOCUMENTDB_LIVE_OPENAI_API_VERSION | none | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:160 | Test-only | Live test only; KeyError if the live test runs without it. |
AZURE_DOCUMENTDB_LIVE_OPENAI_ENDPOINT | none | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:158 | Test-only | Live test only; KeyError if the live test runs without it. |
AZURE_DOCUMENTDB_LIVE_TENANT_ID | none | lib/crewai-tools/tests/tools/test_azure_documentdb_vector_search_e2e.py:171 | Test-only | Needed only for the Azure CLI identity branch of the live test. |
AZURE_ENDPOINT | none | lib/crewai/src/crewai/llms/providers/azure/completion.py:134 | LLM provider | Falls back to AZURE_OPENAI_ENDPOINT, then AZURE_API_BASE. |
AZURE_OPENAI_ENDPOINT | none | lib/crewai/src/crewai/llms/providers/azure/completion.py:135 | LLM provider, Tool | Second endpoint fallback for Azure; the DocumentDB tool (vector_search.py:177) has no embedding endpoint without it. |
AZURE_TENANT_ID | none | lib/crewai-tools/src/crewai_tools/tools/azure_documentdb_vector_search_tool/README.md:75 | Documented only | Not read by CrewAI code; the README example and azure-identity read it. |
BASE_URL | none | lib/crewai/src/crewai/utilities/llm_utils.py:527 | Library | First of four base-URL fallbacks for the default LLM; with none set, the provider default is used. |
BEDROCK_AGENT_ALIAS_ID | none | lib/crewai-tools/src/crewai_tools/aws/bedrock/agents/invoke_agent_tool.py:61 | Tool | The Bedrock agent invoker has no alias unless one is passed. |
BEDROCK_AGENT_ID | none | lib/crewai-tools/src/crewai_tools/aws/bedrock/agents/invoke_agent_tool.py:60 | Tool | The Bedrock agent invoker has no agent unless one is passed. |
BEDROCK_KB_ID | none | lib/crewai-tools/src/crewai_tools/aws/bedrock/knowledge_base/retriever_tool.py:60 | Tool | The knowledge-base retriever has no knowledge base unless one is passed. |
BRAVE_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/brave_search_tool/base.py:126 | Tool | The Brave search tools cannot authenticate; brave_search_tool.py:129 indexes the variable directly and raises KeyError. |
BRIGHTDATA_API_URL | https://api.brightdata.com | lib/crewai-tools/src/crewai_tools/tools/brightdata_tool/brightdata_dataset.py:20 | Tool | The public Bright Data API is used. |
BRIGHTDATA_DEFAULT_POLLING_INTERVAL | 1 | lib/crewai-tools/src/crewai_tools/tools/brightdata_tool/brightdata_dataset.py:23 | Tool | Dataset jobs are polled every second. |
BRIGHTDATA_DEFAULT_TIMEOUT | 600 | lib/crewai-tools/src/crewai_tools/tools/brightdata_tool/brightdata_dataset.py:21 | Tool | Dataset jobs wait up to 600 seconds. |
BRIGHT_DATA_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/brightdata_tool/brightdata_unlocker.py:99 | Tool | The Bright Data tools have no key unless one is passed. |
BRIGHT_DATA_ZONE | none | lib/crewai-tools/src/crewai_tools/tools/brightdata_tool/brightdata_unlocker.py:100 | Tool | The unlocker and SERP tools have no zone unless one is passed. |
BROWSERBASE_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/browserbase_load_tool/browserbase_load_tool.py:16 | Tool | Browserbase and Stagehand (stagehand_tool.py:208) cannot open a remote browser. |
BROWSERBASE_PROJECT_ID | none | lib/crewai-tools/src/crewai_tools/tools/browserbase_load_tool/browserbase_load_tool.py:17 | Tool | As BROWSERBASE_API_KEY. |
CONTAINER_ID | "" | lib/crewai/src/crewai/events/listeners/tracing/utils.py:386 | Library | Read only when HOSTNAME is unset; the machine fingerprint for traces simply omits a container id. |
CREWAI_AGENT_INPUTS | "" (set by the CLI) | lib/cli/src/crewai_cli/run_agent.py:38 | CLI plumbing | Set at line 100 before the agent subprocess starts; empty means the agent runs with no inputs. |
CREWAI_BEDROCK_S3_BUCKET | none | lib/crewai-files/src/crewai_files/uploaders/factory.py:196 | Files | Creating a Bedrock file uploader raises ValueError ("Bedrock file uploads are not configured") unless bucket_name is passed. |
CREWAI_BEDROCK_S3_BUCKET_OWNER | none | lib/crewai-files/src/crewai_files/formatting/api.py:333 | Files | No expected bucket owner is sent with S3 references. |
CREWAI_CLI_RUNNER_PACKAGE_DIR | none (set by the CLI) | lib/cli/src/crewai_cli/run_crew.py:87 | CLI plumbing | Read inside the project subprocess bootstrap; the CLI always sets it, and without it the bootstrap raises KeyError. |
CREWAI_DEPLOYMENT_INSTANCE_UUID | none | lib/crewai-tools/src/crewai_tools/tools/crewai_platform_tools/integrations_client.py:229 | Tool | Platform integration requests carry no deployment-instance header. |
CREWAI_DESERIALIZE_CALLBACKS | "" | lib/crewai/src/crewai/types/callback.py:28 | Library (guard) | Callback references read from checkpoint data are refused rather than imported by dotted path. |
CREWAI_DISABLE_TELEMETRY | unset | lib/crewai-core/src/crewai_core/telemetry.py:357 | Core | Telemetry stays on unless OTEL_SDK_DISABLED or CREWAI_DISABLE_TRACKING is set. |
CREWAI_DISABLE_VERSION_CHECK | "" | lib/crewai/src/crewai/events/utils/console_formatter.py:75 | Library | The console checks PyPI for a newer crewai version (skipped anyway under CI). |
CREWAI_DMN | none | lib/cli/src/crewai_cli/utils.py:102 | CLI | The CLI treats the session as possibly interactive and may prompt. |
CREWAI_ENTERPRISE_ACTION_AUTH_TOKEN | none | lib/crewai/src/crewai/context.py:47 | Library, Tool | Platform action tools have no integration token unless one is set in the run context; the CLI sets it (agent_tools.py:114). |
CREWAI_EVAL_AWAITING_RUN | "" | lib/crewai-core/src/crewai_core/eval_handshake.py:80 | Core, CLI plumbing | No evaluation is waiting on this run. |
CREWAI_FACTORY | false | lib/crewai-tools/src/crewai_tools/tools/crewai_platform_tools/integrations_client.py:161 | Tool | TLS certificates are verified on platform integration calls. Set to true, it turns verification off on all four call sites (lines 161, 199, 256, 306). |
CREWAI_JSON_CREW_DEFINITION | none | lib/cli/src/crewai_cli/run_crew.py:128 | CLI plumbing | The JSON crew runner uses its default crew file. |
CREWAI_JSON_CREW_INPUTS | none | lib/cli/src/crewai_cli/run_crew.py:130 | CLI plumbing | The JSON crew runs without inputs. |
CREWAI_NL2SQL_ALLOW_DML | "" | lib/crewai-tools/src/crewai_tools/tools/nl2sql/nl2sql_tool.py:272 | Tool (guard) | Write statements are refused unless the tool was built with allow_dml. |
CREWAI_ORGANIZATION_UUID | none | lib/crewai/src/crewai/skills/registry.py:254 | Library | The skills registry client is built without an organisation. |
CREWAI_ORG_ID | "" | lib/crewai/src/crewai/events/listeners/tracing/trace_listener.py:204 | Library, CLI | Trace context carries no organisation id; crewai traces status shows it unset (cli.py:1415). |
CREWAI_PASTED_API_KEYS | "" | lib/cli/src/crewai_cli/api_keys.py:308 | CLI plumbing | No keys were pasted at a prompt in this session; the CLI clears it at startup (cli.py:136). |
CREWAI_PERSONAL_ACCESS_TOKEN | none | lib/crewai-tools/src/crewai_tools/tools/generate_crewai_automation_tool/generate_crewai_automation_tool.py:32 | Tool | The automation generator tool cannot authenticate to the platform unless a token is passed. |
CREWAI_PLATFORM_INTEGRATION_TOKEN | "" | lib/crewai-tools/src/crewai_tools/tools/crewai_platform_tools/misc.py:8 | Tool | Platform tools send an empty bearer token and the platform rejects the call. |
CREWAI_PLUS_URL | https://app.crewai.com | lib/crewai-core/src/crewai_core/plus_api.py:291 | Core, CLI, Tool | The URL saved in settings.json is used, then https://app.crewai.com. |
CREWAI_RUNNER_SOURCE_DIR | none | lib/cli/src/crewai_cli/run_crew.py:66 | CLI plumbing | No local crewai source checkout is put first on the subprocess path; the installed package is used. |
CREWAI_STORAGE_DIR | name of the current directory | lib/crewai-core/src/crewai_core/paths.py:13 | Core | Stores go under the user data directory in a folder named after the working directory, so two projects with the same folder name share stores. |
CREWAI_TESTING | "" | lib/crewai/src/crewai/events/listeners/tracing/utils.py:232 | Library (test switch) | Normal behaviour: tracing prompts may show. conftest.py:277 sets it to true for the suite. |
CREWAI_TRACING_ENABLED | "" | lib/crewai/src/crewai/events/listeners/tracing/utils.py:168 | Library, CLI | The user's saved choice from crewai traces enable or disable decides; false or 0 turns tracing off, true or 1 turns it on. |
CREWAI_TRAINED_AGENTS_FILE_ENV | trained_agents_data.pkl (stands for CREWAI_TRAINED_AGENTS_FILE) | lib/crewai/src/crewai/agent/core.py:1553 | Library (constant name) | Agents load trained suggestions from trained_agents_data.pkl in the working directory, if present. |
CREWAI_USER_ID | anonymous | lib/crewai/src/crewai/events/listeners/tracing/trace_listener.py:203 | Library | Traces are attributed to "anonymous". |
CREWAI_USER_PAT | none | lib/crewai/src/crewai/skills/registry.py:251 | Library | Falls back to the platform integration token, then the saved CLI login token. |
CREW_AWS_ACCESS_KEY_ID | none | lib/crewai-tools/src/crewai_tools/aws/s3/reader_tool.py:37 | Tool | The S3 reader and writer tools pass no key, leaving boto3's credential chain to decide. |
CREW_AWS_REGION | us-east-1 | lib/crewai-tools/src/crewai_tools/aws/s3/reader_tool.py:36 | Tool | S3 tools use us-east-1. |
CREW_AWS_SEC_ACCESS_KEY | none | lib/crewai-tools/src/crewai_tools/aws/s3/reader_tool.py:38 | Tool | As CREW_AWS_ACCESS_KEY_ID. |
DAYTONA_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/daytona_sandbox_tool/daytona_base_tool.py:35 | Tool | The Daytona sandbox tools cannot create a sandbox unless a key is passed. |
DAYTONA_API_URL | none | lib/crewai-tools/src/crewai_tools/tools/daytona_sandbox_tool/daytona_base_tool.py:40 | Tool | The Daytona SDK's default URL is used. |
DAYTONA_TARGET | none | lib/crewai-tools/src/crewai_tools/tools/daytona_sandbox_tool/daytona_base_tool.py:45 | Tool | The Daytona SDK's default target is used. |
DECL_FLOW_ENV_PROBE | none | lib/cli/tests/test_run_declarative_flow.py:378 | Test-only | A probe the test writes to a .env file to check it is loaded. |
E2B_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/e2b_sandbox_tool/e2b_base_tool.py:36 | Tool | The E2B sandbox tools cannot create a sandbox unless a key is passed. |
E2B_DOMAIN | none | lib/crewai-tools/src/crewai_tools/tools/e2b_sandbox_tool/e2b_base_tool.py:43 | Tool | The E2B SDK's default domain is used. |
ENTERPRISE_CREWAI_DEP_PATH | none | lib/devtools/src/crewai_devtools/cli.py:1312 | Devtools | The release tool skips or cannot locate the enterprise dependency path to bump. |
ENTERPRISE_EXTRA_PACKAGES | "" | lib/devtools/src/crewai_devtools/cli.py:1315 | Devtools | No extra enterprise packages are bumped. |
ENTERPRISE_REPO | none | lib/devtools/src/crewai_devtools/cli.py:1308 | Devtools | The release tool has no enterprise repository to update. |
ENTERPRISE_VERSION_DIRS | "" | lib/devtools/src/crewai_devtools/cli.py:1310 | Devtools | No enterprise version directories are bumped. |
ENTERPRISE_WORKFLOW_PATHS | "" | lib/devtools/src/crewai_devtools/cli.py:1320 | Devtools | No enterprise workflow files are updated. |
EVAL_TOKEN_ENV | "" (stands for CREWAI_EVAL_AWAITING_RUN) | lib/cli/src/crewai_cli/run_crew.py:913,933 | CLI (constant name) | As CREWAI_EVAL_AWAITING_RUN: nothing to pass on to the project subprocess. |
EVENT_ACTION | none | .github/workflows/ftc-require-issue.yml:111 | CI-only | Not treated as a reopen, so the reopen-by-maintainer exemption does not apply. |
EXA_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/exa_tools/exa_search_tool.py:58 | Tool | The Exa tool has no key unless one is passed. |
EXA_BASE_URL | none | lib/crewai-tools/src/crewai_tools/tools/exa_tools/exa_search_tool.py:63 | Tool | The Exa SDK's default URL is used. |
FIRECRAWL_API_KEY | none ("" in the script) | lib/crewai-tools/src/crewai_tools/tools/firecrawl_search_tool/firecrawl_search_tool.py:69 | Tool, Script | The three Firecrawl tools declare it as their required key; scripts/wharf_runner.py:70 sends an empty key to the Firecrawl MCP server. |
FROM_THE_PROJECT | none | lib/cli/tests/test_run_crew.py:1835 | Test-only | A probe the test writes to a project .env to check it reaches the subprocess. |
GEMINI_API_KEY | none | lib/crewai/src/crewai/llms/providers/gemini/completion.py:53 | LLM provider | Second choice after GOOGLE_API_KEY; with neither, Gemini needs Vertex AI credentials. |
GITHUB_ACTIONS | none | conftest.py:466 | Test-only | The cassette configuration takes its local, not CI, branch. |
GOOGLE_API_KEY | none | lib/crewai/src/crewai/llms/providers/gemini/completion.py:53 | LLM provider, Files | Falls back to GEMINI_API_KEY; the Gemini file uploader (crewai_files uploaders/gemini.py:108) has no key. |
GOOGLE_CLOUD_LOCATION | none | lib/crewai/src/crewai/llms/providers/gemini/completion.py:108 | LLM provider | Vertex AI runs without an explicit location. |
GOOGLE_CLOUD_PROJECT | none | lib/crewai/src/crewai/llms/providers/gemini/completion.py:106 | LLM provider | Vertex AI mode is not selected from a project id. |
GOOGLE_GENAI_USE_VERTEXAI | "" | lib/crewai/src/crewai/llms/providers/gemini/completion.py:113 | LLM provider | The Gemini API, not Vertex AI, is used. |
HOST | none | lib/cli/tests/test_api_keys.py:533 | Test-only | Appears inside generated project source in a test fixture; no product code reads it. |
HTTPS_PROXY | none | lib/crewai-core/src/crewai_core/plus_api.py:160-176 | Core, CLI | Direct connections. Platform requests take the proxy from the machine environment captured at import, never from a project .env. |
HYPERBROWSER_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/hyperbrowser_load_tool/hyperbrowser_load_tool.py:49 | Tool | The Hyperbrowser tool has no key unless one is passed. |
LINKUP_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/linkup/linkup_search_tool.py:54 | Tool | The Linkup tool has no key unless one is passed. |
LITELLM_FAILURE_CALLBACKS | "" | lib/crewai/src/crewai/llm.py:2716 | Library | No LiteLLM failure callbacks are registered. |
LITELLM_SUCCESS_CALLBACKS | "" | lib/crewai/src/crewai/llm.py:2709 | Library | No LiteLLM success callbacks are registered. |
LOCALAPPDATA | none | lib/crewai-core/src/crewai_core/token_manager.py:88 | Core (Windows) | On Windows, the CLI has no base directory for its encrypted token store. |
MANAGED_AGENT_TEST_CONFIG | none | lib/cli/tests/test_run_agent.py:76 | Test-only | A probe the test writes to a project .env. |
MANAGED_AGENT_VALIDATION_ENV | none | lib/cli/tests/test_create_agent.py:170 | Test-only | A probe the test writes to a project .env. |
MINDS_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/ai_mind_tool/ai_mind_tool.py:47 | Tool | The AI Minds tool has no key unless one is passed. |
MODEL | none (template: openai/gpt-6-luna) | lib/crewai/src/crewai/utilities/llm_utils.py:501 | Library | Falls back to MODEL_NAME, then OPENAI_MODEL_NAME, then the library default model. |
MODEL_NAME | none | lib/crewai/src/crewai/utilities/llm_utils.py:502 | Library | Second model-name fallback. |
MULTION_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/multion_tool/multion_tool.py:52 | Tool | The MultiOn tool has no key unless one is passed. |
OLLAMA_API_BASE | http://localhost:11434 | lib/cli/src/crewai_cli/model_catalog.py:302 | CLI | The model picker asks the local Ollama server on port 11434. |
OLLAMA_HOST | http://localhost:11434 | lib/cli/src/crewai_cli/model_catalog.py:304 | CLI, LLM provider | As above; also the base-URL variable of the Ollama entry in openai_compatible/completion.py:64. |
OPENAI_API_BASE | none | lib/crewai/src/crewai/llms/providers/openai/completion.py:430 | LLM provider | Second choice after OPENAI_BASE_URL; with neither, the OpenAI default endpoint. |
OPENAI_API_KEY | none | lib/crewai/src/crewai/llms/providers/openai/completion.py:568 | LLM provider, Tool | Building an OpenAI client without an explicit key raises MissingAPIKeyError; several tools also use it for embeddings. A CI secret in agents-md-eval.yml. |
OPENAI_BASE_URL | none | lib/crewai/src/crewai/llms/providers/openai/completion.py:429 | LLM provider | Falls back to OPENAI_API_BASE, then the OpenAI default. |
OPENAI_MODEL_NAME | none | lib/crewai/src/crewai/utilities/llm_utils.py:503 | Library | Third model-name fallback; then the library default. |
OPENROUTER_API_KEY | none | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:50 | LLM provider | The OpenRouter provider and embedder (openrouter_provider.py:31) have no key unless one is passed. |
OPEN_ROUTER_API_KEY | none | lib/crewai/tests/test_llm.py:587 | Test-only | The OpenRouter live test is skipped. |
ORACLE_DB_DSN | none | lib/crewai-tools/src/crewai_tools/tools/oracle_vector_search_tool/README.md:38 | Documented only | Not read by code; the README example reads it to build a connection. |
ORACLE_DB_SECRET | none | lib/crewai-tools/src/crewai_tools/tools/oracle_vector_search_tool/README.md:37 | Documented only | As ORACLE_DB_DSN. |
ORACLE_DB_USERNAME | none | lib/crewai-tools/src/crewai_tools/tools/oracle_vector_search_tool/README.md:36 | Documented only | As ORACLE_DB_DSN. |
OTEL_SDK_DISABLED | unset | lib/crewai-core/src/crewai_core/telemetry.py:356 | Core | Telemetry stays on unless another opt-out is set. The OpenTelemetry SDK reads it too. |
OXYLABS_PASSWORD | none | lib/crewai-tools/src/crewai_tools/tools/oxylabs_base_tool/oxylabs_base_tool.py:181 | Tool | The Oxylabs tools have no credentials unless they are passed. |
OXYLABS_USERNAME | none | lib/crewai-tools/src/crewai_tools/tools/oxylabs_base_tool/oxylabs_base_tool.py:180 | Tool | As OXYLABS_PASSWORD. |
PARALLEL_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/parallel_tools/parallel_search_tool.py:78 | Tool | The Parallel search tool has no key unless one is passed. |
PASTED_KEYS_ENV | "" (stands for CREWAI_PASTED_API_KEYS) | lib/cli/src/crewai_cli/api_keys.py:42 | CLI (constant name) | As CREWAI_PASTED_API_KEYS. |
PATRONUS_API_KEY | none ("" in the predefined-criteria tool) | lib/crewai-tools/src/crewai_tools/tools/patronus_eval_tool/patronus_eval_tool.py:43 | Tool | patronus_eval_tool.py indexes the variable directly and raises KeyError; the predefined-criteria tool sends an empty key. |
PR_NUMBER | none (set by the workflow) | .github/workflows/ftc-require-issue.yml:46 | CI-only | KeyError in the check step. |
PYTEST_VCR_RECORD_MODE | once | conftest.py:458 | Test-only | Cassettes are recorded once and replayed after. |
REDIS_URL | none | lib/crewai-core/src/crewai_core/lock_store.py:33 | Core, Library (A2A) | Locks are file locks in the system temp directory instead of Redis locks, and the A2A task cache (a2a/utils/task.py:102) stays in memory. |
RELEASE_TAG | the package version | .github/workflows/publish.yml:118 | CI-only | The release notification links the tag named after the version. |
REPO | none (set by the workflow) | .github/workflows/ftc-require-issue.yml:45 | CI-only | KeyError in the check step. |
RUN_BEDROCK_LIVE_TESTS | false | lib/crewai/tests/agents/test_native_tool_calling.py:800 | Test-only | The live Bedrock tool-calling test is skipped. |
RUN_COUNTED_ENV | unset (stands for CREWAI_RUN_COUNTED) | lib/cli/src/crewai_cli/run_crew.py:721 | CLI (constant name) | The project subprocess counts its own run; the CLI sets it to 1 when it has already counted (run_declarative_flow.py:587). |
SCRAPEGRAPH_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/scrapegraph_scrape_tool/scrapegraph_scrape_tool.py:113 | Tool | The ScrapeGraph tool has no key unless one is passed. |
SCRAPFLY_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/scrapfly_scrape_website_tool/scrapfly_scrape_website_tool.py:66 | Tool | The Scrapfly tool has no key unless one is passed. |
SERPAPI_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/serpapi_tool/serpapi_base_tool.py:44 | Tool | The SerpApi tools have no key unless one is passed. |
SERPER_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/serper_dev_tool/serper_dev_tool.py:251 | Tool | The Serper search tool indexes the variable directly and raises KeyError when it runs. |
SERPLY_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/serply_api_tool/serply_web_search_tool.py:67 | Tool | The five Serply tools index the variable directly and raise KeyError. |
SINGLESTOREDB_URL | none | lib/crewai-tools/src/crewai_tools/tools/singlestore_search_tool/singlestore_search_tool.py:53 | Tool | Declared as the tool's connection variable; without it the connection comes from SINGLESTOREDB_HOST or arguments. |
SNOWFLAKE_ACCOUNT | none | lib/crewai/src/crewai/llms/providers/snowflake/completion.py:122 | LLM provider | Second account fallback after SNOWFLAKE_ACCOUNT_URL. |
SNOWFLAKE_ACCOUNT_ID | none | lib/crewai/src/crewai/llms/providers/snowflake/completion.py:123 | LLM provider | Third account fallback. |
SNOWFLAKE_ACCOUNT_IDENTIFIER | none | lib/crewai/src/crewai/llms/providers/snowflake/completion.py:124 | LLM provider | Fourth account fallback; with none, the Snowflake client has no account URL. |
SNOWFLAKE_ACCOUNT_URL | none | lib/crewai/src/crewai/llms/providers/snowflake/completion.py:114 | LLM provider | The URL is built from an account name instead. |
SSL_CERT_FILE | none (the CLI sets certifi's bundle) | lib/crewai-core/src/crewai_core/plus_api.py:232 | Core, CLI | Platform requests use the default CA store; the CLI's provider download overwrites it with certifi's bundle (provider.py:162). |
TAVILY_API_KEY | none | lib/crewai-tools/src/crewai_tools/tools/tavily_search_tool/tavily_search_tool.py:61 | Tool | The four Tavily tools have no key unless one is passed. |
TOKENIZERS_PARALLELISM | false (written, not read) | lib/crewai/src/crewai/knowledge/knowledge.py:66 | Library | Not read: importing the knowledge module sets it to false for the Hugging Face tokenizers library. |
TRACING_ENV_VAR | (stands for CREWAI_TRACING_ENABLED) | lib/cli/src/crewai_cli/experimental/eval_crew.py:77 | CLI (constant name) | As CREWAI_TRACING_ENABLED; crewai eval refuses a project where it is false because evaluation needs the trace. |
UV_RUN_RECURSION_DEPTH | none | lib/cli/src/crewai_cli/run_declarative_flow.py:612 | CLI | The CLI concludes it is not already inside uv run. |
ZAPIER_API_KEY | none | lib/crewai-tools/src/crewai_tools/adapters/zapier_adapter.py:62 | Tool | The Zapier adapter and action tool have no key unless one is passed. |
CREWAI_DISABLE_TRACKING | unset | lib/crewai-core/src/crewai_core/telemetry.py:358 | Core | Telemetry stays on unless another opt-out is set. Not in the evidence list; added because the source reads it. |
The evidence scan matched only variables named as string literals in an accessor call or a .env sample, so it missed variables read through a tuple, a prefix test or an SDK. Reading the source turned up twelve more, in the ten rows below (one row covers three proxy variables). They are listed separately so the evidence count above stays checkable. The source therefore reads at least 163 variables, the 151 in the inventory above plus these 12, and this section carries 161 variable rows.
| Variable | Default | Read at | Scope | When absent |
|---|---|---|---|---|
AWS_SESSION_TOKEN | none | lib/crewai/src/crewai/llms/providers/bedrock/completion.py:306 | LLM provider | No session token; long-lived keys or boto3's chain are used. |
CI | "" | lib/crewai/src/crewai/events/utils/console_formatter.py:72 | Library | The console runs its PyPI version check (unless disabled). |
HOSTNAME | none | lib/crewai/src/crewai/events/listeners/tracing/utils.py:385 | Library | Falls back to CONTAINER_ID for the trace fingerprint. |
SSL_CERT_DIR | none | lib/crewai-core/src/crewai_core/plus_api.py:234 | Core | Platform requests use the default CA store. |
HTTP_PROXY, ALL_PROXY, NO_PROXY | none | lib/crewai-core/src/crewai_core/plus_api.py:160 | Core, CLI | Any variable ending in _PROXY is captured from the machine environment at import; absent, connections are direct. |
UV_ENV_FILE | none | lib/crewai-core/src/crewai_core/plus_api.py:164 | Core | Treated as a transport variable because uv loads that file before Python starts. |
DATABRICKS_HOST | none | lib/crewai-tools/src/crewai_tools/tools/databricks_query_tool/databricks_query_tool.py:147 | Tool | The Databricks tool falls back to a configured profile. |
DATABRICKS_TOKEN | none | lib/crewai-tools/src/crewai_tools/tools/databricks_query_tool/databricks_query_tool.py:147 | Tool | As DATABRICKS_HOST. |
SINGLESTOREDB_HOST | none | lib/crewai-tools/src/crewai_tools/tools/singlestore_search_tool/singlestore_search_tool.py:62 | Tool | The tool needs SINGLESTOREDB_URL or explicit arguments. |
KUBERNETES_SERVICE_HOST | none | lib/crewai-core/src/crewai_core/runtime_env.py:147 | Core | A presence marker for the runtime context telemetry reports; absent, a container is recognised only by /.dockerenv. runtime_env.py checks dozens of similar presence markers (CI, serverless, PaaS, hosted IDE, notebook), not itemised here. |
| Function | Where | What it does |
|---|---|---|
_is_telemetry_disabled | lib/crewai-core/src/crewai_core/telemetry.py:354 | Returns true when any of OTEL_SDK_DISABLED, CREWAI_DISABLE_TELEMETRY or CREWAI_DISABLE_TRACKING is set to a truthy value. Every telemetry call checks it first. |
_env_flag_enabled | lib/crewai-core/src/crewai_core/telemetry.py:327 | Reads one variable as a boolean, accepting true/1/yes/on and false/0/no/off. An unrecognised value logs one warning and counts as unset. |
_sets_transport | lib/crewai-core/src/crewai_core/plus_api.py:153 | Decides whether a variable affects the proxy or CA bundle of platform requests. Those variables are frozen from the machine environment at import so a project's .env cannot redirect the saved login. |
load_env_vars | lib/cli/src/crewai_cli/utils.py:257 | Reads KEY=VALUE pairs from a project folder's .env file into a dictionary. |
write_env_file | lib/cli/src/crewai_cli/utils.py:270 | Writes the API keys collected by crewai create to the project's .env with owner-only permissions. It refuses to write through a symlink. |
_get_or_create_key | lib/crewai-core/src/crewai_core/token_manager.py:32 | Loads the Fernet key from secret.key in the credential directory, or generates and writes a new one there. |
save_tokens | lib/crewai-core/src/crewai_core/token_manager.py:50 | Encrypts the access token and its expiry time and writes them atomically to tokens.enc. |
get_token | lib/crewai-core/src/crewai_core/token_manager.py:65 | Decrypts tokens.enc and returns the access token, or nothing if it is missing or expired. |
Where secrets come from. Model and tool keys are read from the process environment or passed as constructor arguments; the library never stores them. crewai create prompts for the provider's keys and writes them to the new project's .env with owner-only permissions (write_env_file), and crewai run loads that file into the subprocess it starts. The repository holds no .env.example or other sample file and no single configuration reference: the variables a project needs are described per tool in each tool's README and in crewai_cli/constants.py's ENV_VARS prompt table, which covers model providers only.
The CLI login. crewai login runs the OAuth2 device flow and stores the access token Fernet-encrypted at rest in tokens.enc, with the key generated and kept on the same disk; section 8 ("Token storage") describes the key file, its permissions and what the encryption does and does not protect against. Organisation and platform settings live in ~/.config/crewai/settings.json, written atomically with mode 0600 into a directory tightened to 0700.
CI secrets. GitHub Actions secrets reach workflows as environment variables: PYPI_API_TOKEN as UV_PUBLISH_TOKEN in publish.yml:84 and nightly.yml:120, ANTHROPIC_API_KEY and OPENAI_API_KEY in agents-md-eval.yml, and GITHUB_TOKEN in the PR housekeeping workflows. The test workflow uses a fake OPENAI_API_KEY and replays recorded cassettes.
Rotation. Nothing in the repository rotates a secret. The only lifetime handled in code is the CLI token's expiry: get_token returns nothing once the stored expiry passes and the user must log in again. Provider and tool keys live as long as the user keeps them in the environment.
Validation. Nothing validates configuration at startup. A missing provider key surfaces when that provider's client is built (Anthropic and OpenAI raise MissingAPIKeyError), and a missing tool key surfaces when the tool is constructed or first called, some as a bare KeyError. The exceptions are the opt-in guards, which fail safe: CREWAI_DESERIALIZE_CALLBACKS and CREWAI_NL2SQL_ALLOW_DML keep the risky path off when unset. One switch goes the other way: CREWAI_FACTORY=true turns off TLS verification on all platform integration calls.
Fair. Configuration is externalised (at least 163 environment variables: the 151 in the evidence inventory and 12 more found in the source, a pydantic settings file for the CLI, CI secrets through GitHub Actions) and the defaults are sensible. But reads are scattered through os.environ across hundreds of files with no central schema, and there is no .env.example or single configuration reference in the repository. A missing provider key surfaces only when the client is first built, and nothing rotates secrets beyond the CLI token's expiry. The technical assessment gives the reasoning per criterion.
The repository deploys no runtime infrastructure. There is no Dockerfile, compose file, IaC, Procfile, Makefile or .env.example: of the 26 probed infrastructure paths only .github/workflows and pyproject.toml are present, and the product ships as PyPI packages. The repository is a uv workspace of six packages, all at version 1.15.41: crewai, crewai-core, crewai-tools, crewai-files, crewai-cli and crewai-devtools. Five of them are published; crewai-devtools is skipped by the publish loop as a private package. Once installed, the framework runs inside the developer's own Python process (supported Python is >=3.10,<3.14), so "production" is wherever the developer's application runs, and none of that is described here.
Four environments appear in the code. Only GitHub Actions is configured by this repository; the developer machine is whatever runs the installed package, and CrewAI AMP is a hosted platform that the CLI talks to but whose infrastructure lives elsewhere.
| Environment | What runs there | Shared with others | Isolated | How code reaches it |
|---|---|---|---|---|
| Developer machine | The user's project, the installed crewai packages, the crewai CLI, and the local stores of section 11 under appdirs.user_data_dir. | The same PyPI artefacts every other user installs; provider API keys from the user's own environment. | Local stores are keyed by the project directory name (or CREWAI_STORAGE_DIR), so two projects on one machine do not share flow state, memory or knowledge. | uv add crewai or pip install from PyPI; crewai run spawns uv run subprocesses inside the project. |
| GitHub Actions | 19 workflows: tests on Python 3.10 to 3.13 in 8 groups each with network blocked, ruff, mypy, CodeQL, pip-audit, docs link checks, repository housekeeping, and the two publish workflows. | The repository, a uv cache keyed on uv.lock, and stored test durations. | Secrets are scoped to GitHub environments: pypi holds the publish token, agents-md-eval holds model API keys and runs only for same-repository branches. | Pull requests, pushes to main, schedules and manual dispatch (table below). |
| PyPI | The five published packages, plus a nightly .devYYYYMMDD pre-release of each when main has changed. | Every user of the framework. | Nothing environment-specific: one release serves all users. | publish.yml (manual, optional release tag) and nightly.yml (daily at 06:00 UTC), both via uv publish with the PYPI_API_TOKEN secret. |
| CrewAI AMP | Deployed crews and flows, the execution trace viewer, organisations and the tool repository. Hosted at https://app.crewai.com by default; outside this repository. | The platform account and organisation the CLI is logged in to. | Each deployment is a separate AMP crew identified by UUID or project name. | crewai deploy create and crewai deploy push: AMP redeploys from the user's git remote, or the CLI uploads a ZIP of the project together with the variables from its .env. |
A release is a manual dispatch of publish.yml: it checks out the given tag (or the current ref), runs uv build --all-packages, hands the dist/ artefact to a second job in the pypi GitHub environment, publishes every wheel and sdist except crewai_devtools, and posts a Slack message built from the GitHub release notes. That job is granted id-token: write, but the upload authenticates with the PYPI_API_TOKEN secret. nightly.yml does the same each day at 06:00 UTC when main has a commit in the last 25 hours, after stamping every package and every cross-package pin with BASE.devYYYYMMDD; its uv publish --check-url skips files already on PyPI, so a same-day re-run is harmless. The maintainer console scripts bump-version, tag and release in crewai-devtools (section 3) prepare the version and tag before a dispatch.
A deploy to CrewAI AMP starts from the developer machine. deploy runs the pre-deploy validation, asks AMP how it holds the deployment's code, and then either asks AMP to redeploy from its git source (deploy_by_uuid or deploy_by_name) or builds a ZIP of the project with create_project_zip and uploads it through update_crew_from_zip together with the .env variables that fetch_and_json_env_file reads. Everything after the upload (build, hosting, scaling) happens inside AMP and is not visible here.
| Function | Where | What it does |
|---|---|---|
deploy | lib/cli/src/crewai_cli/deploy/main.py:432 | Validates the local project, opens a deployment telemetry span, and redeploys the crew on AMP either from AMP's own source or from the local git remote or ZIP. It then prints the deployment details and opens the deployment page. |
create_project_zip | lib/cli/src/crewai_cli/deploy/archive.py:48 | Collects the project's deployable files (from git when there is a repository, otherwise by walking the directory), stages them and writes a ZIP archive. It raises an error and leaves no partial archive when nothing deployable is found. |
update_crew_from_zip | lib/crewai-core/src/crewai_core/plus_api.py:556 | Uploads a project ZIP and its environment variables to an existing AMP deployment with a POST to the crews API. |
deploy_by_uuid | lib/crewai-core/src/crewai_core/plus_api.py:475 | Asks AMP to redeploy the crew with the given UUID from the source AMP already holds, optionally forcing it while a model comparison runs. |
deploy_by_name | lib/crewai-core/src/crewai_core/plus_api.py:460 | Asks AMP to redeploy the crew with the given project name, sending the project id from [tool.crewai] when one is set. |
fetch_and_json_env_file | lib/cli/src/crewai_cli/utils.py:198 | Reads the project's .env file into a dictionary of key and value pairs, skipping blank and comment lines. It prints an error when the file is missing. |
db_storage_path | lib/crewai-core/src/crewai_core/paths.py:16 | Returns, and creates if needed, the per-project data directory under appdirs.user_data_dir, named after the current directory or CREWAI_STORAGE_DIR. |
The evidence probed 26 paths that would describe infrastructure, a pipeline or a runtime. Two are present. The absences are the description: no container image, no serverless or IaC template, no orchestration manifest and no environment sample, because the deliverable is a set of Python packages rather than a running service.
| Path | Present | What that means here |
|---|---|---|
| Dockerfile | no | No container image is built or shipped. |
| docker-compose.yml | no | No local multi-service stack; tests that need a database start their own container or are skipped. |
| compose.yml | no | As above. |
| serverless.yml | no | No serverless deployment. |
| template.yml | no | No AWS SAM or CloudFormation template. |
| template.yaml | no | As above. |
| samconfig.toml | no | As above. |
| main.tf | no | No Terraform. |
| terraform | no | As above. |
| cdk.json | no | No AWS CDK app. |
| Procfile | no | No process definition for a PaaS. |
| Makefile | no | Build and test commands live in the workflows and in uv, not in make targets. |
| k8s | no | No Kubernetes manifests. |
| .github/workflows | yes | The only pipeline: 19 GitHub Actions workflows, listed below. |
| .gitlab-ci.yml | no | No GitLab pipeline. |
| Jenkinsfile | no | No Jenkins pipeline. |
| azure-pipelines.yml | no | No Azure DevOps pipeline. |
| .env.example | no | No environment sample; variables are documented per feature (section 9). |
| .env.sample | no | As above. |
| composer.json | no | No PHP code. |
| package.json | no | No Node package at the root; the flow visualisation's JavaScript is a static asset inside the crewai package. |
| pyproject.toml | yes | The uv workspace root: members under lib/, ruff, mypy and pytest settings, and the exclude-newer = "3 days" resolution cooldown. |
| go.mod | no | No Go code. |
| Cargo.toml | no | No Rust code. |
| pom.xml | no | No Java code. |
| .gitattributes | no | No line-ending or LFS rules. |
The 19 workflows fall into four groups: quality gates on pull requests (tests, lint, types, CodeQL, pip-audit, docs links, AGENTS.md guards), release (publish and nightly), cache and spec maintenance, and contributor housekeeping.
| Workflow | Trigger | What it does |
|---|---|---|
| tests.yml | pull request | Runs pytest on Python 3.10, 3.11, 3.12 and 3.13, each split into 8 groups by stored durations, with network blocked, 60 s timeouts and random order. |
| linter.yml | pull request | When code changed, runs ruff check and ruff format --check over lib/. |
| type-checker.yml | pull request | Runs mypy (strict, with the pydantic plugin) over lib/. |
| codeql.yml | pull request and push to main, ignoring CLI templates | Runs GitHub CodeQL static analysis. |
| vulnerability-scan.yml | pull request, push to main, weekly Monday 09:00 UTC | Installs every group and extra and runs pip-audit over the resolved dependencies, uploading the report. |
| docs-broken-links.yml | pull request or push to main touching docs/, manual | Prunes all but the edge and default doc versions, then runs mint broken-links over the documentation site. |
| agents-md.yml | pull request to main touching CLI templates, the eval or linked docs | Runs the AGENTS.md guard tests in the CLI suite and the eval's own tests. |
| agents-md-eval.yml | pull_request_target to main touching the template AGENTS.md or CLAUDE.md | For same-repository branches only, runs the devtools AGENTS.md eval against four coding-agent models using API keys held in the agents-md-eval environment. |
| publish.yml | manual dispatch, optional release tag | Builds all packages and publishes all but crewai-devtools to PyPI with PYPI_API_TOKEN, then posts the release to Slack. |
| nightly.yml | daily 06:00 UTC, manual | When main changed in the last 25 hours, stamps a .devYYYYMMDD version on every package and pin, builds, and publishes idempotently to PyPI. |
| build-uv-cache.yml | push to main touching uv.lock or pyproject.toml, every 5 days, manual | Installs every dependency group and extra for Python 3.10 to 3.13 and saves the uv cache the other workflows restore, so it does not expire between runs. |
| update-test-durations.yml | push to main touching test files, manual | Runs the full suite with --store-durations and caches the durations the test matrix uses to balance its groups. |
| generate-tool-specs.yml | pull request to main touching crewai-tools, manual | Regenerates lib/crewai-tools/tool.specs.json and commits it back to the branch when it changed. |
| pr-title.yml | pull request opened, edited, synchronised or reopened | Checks for a semantic (conventional commit) pull request title; it runs with continue-on-error, so a bad title is reported without failing the workflow. |
| pr-size.yml | pull request opened, synchronised or reopened | Labels the pull request by diff size. |
| ftc-require-issue.yml | pull_request_target opened, edited or reopened | For first-time contributors (authors who are not members, collaborators or past contributors), requires the pull request to reference an open issue; authors with write access are exempt. |
| first-contributor-welcome.yml | pull_request_target closed | Comments once to welcome a first-time contributor when their pull request is merged. |
| close-associated-prs.yml | issue closed | When an issue is closed as completed by someone with triage or higher permission, closes the open pull requests associated with it. |
| stale.yml | daily 12:10 UTC, manual | Marks issues stale after 30 days and closes them 5 days later; marks pull requests stale after 45 days without closing them. |
CrewAI has no server and no shared database. Every store is local to the machine running the developer's process, single-purpose, and created on first use. Nine stores: flow state SQLite (flow_states.db), kickoff task outputs SQLite (latest_kickoff_task_outputs.db), checkpoint store (JSON files under ./.checkpoints by default, a SQLite checkpoints table when CheckpointConfig is given the SQLite provider), memory store (LanceDB table memories by default, Qdrant Edge optional), knowledge store (ChromaDB collections named knowledge_<collection_name>, or just knowledge when no collection name is given), CLI settings file, CLI token store, training pickles (training_data.pkl, trained_agents_data.pkl), and file upload cache (aiocache, in memory with TTL).
SQLite stores live under appdirs.user_data_dir keyed by the project directory name or CREWAI_STORAGE_DIR, run in WAL mode with busy timeouts. The LanceDB and ChromaDB directories sit under the same per-project directory. Writes that several processes could race on are also serialised by a named lock from crewai_core.lock_store: a file lock in the system temp directory, or a Redis lock when REDIS_URL is set and the redis package is installed.
The checkpoint store is not chosen by the location's suffix. CheckpointConfig uses JsonProvider unless its provider field is set to SqliteProvider; with the SQLite provider, a location without a suffix gets .db appended (lib/crewai/src/crewai/state/checkpoint_config.py:177-201). On restore, detect_provider picks SQLite by reading the file's SQLite magic bytes, not its name (lib/crewai/src/crewai/state/provider/utils.py:11-33). The knowledge collection name is built in lib/crewai/src/crewai/knowledge/storage/knowledge_storage.py:71-75.
A tenth small local file exists beside the nine: .crewai_user.json in the per-project data directory, holding the trace consent flags written by crewai traces enable/disable and the first-run prompt (lib/crewai-core/src/crewai_core/user_data.py:23-26; lib/cli/src/crewai_cli/cli.py:1463,1484). It is listed in the files table below but not counted as a store.
| Store | Technology and location | What it holds | Written and read by | Retention |
|---|---|---|---|---|
| Flow state store | SQLite flow_states.db under the per-project data directory; tables flow_states and pending_feedback. The default behind default_flow_persistence; replaceable through set_flow_persistence_factory. | A JSON snapshot of a flow's state after each persisted method, and one paused-for-feedback record per flow with its context and state. | Flows decorated with @persist and the human feedback pause; read back when a flow is restored by id, forked, or resumed with feedback. | None. Every snapshot is kept; only the pending feedback row is deleted when the flow resumes. |
| Kickoff task outputs store | SQLite latest_kickoff_task_outputs.db under the per-project data directory; table latest_kickoff_task_outputs. | For the latest crew kickoff: each task's id, key, expected output, output, index, inputs and whether it was replayed. | Crew through TaskOutputStorageHandler after each task; read by crew replay and by crewai log-tasks-outputs. | Overwritten: the table is cleared at the start of each kickoff, so it only ever holds the last run. |
| Checkpoint store | JSON files under ./.checkpoints/<branch>/ (relative to the working directory) by default, or a SQLite checkpoints table with the SQLite provider (chosen by the provider field, not the file suffix). | The serialised runtime state of a crew, flow or agent at each configured event (by default task_completed), with branch and parent lineage. | The checkpoint listener on the event bus; read by restore and fork, and by the crewai checkpoint commands. | Keeps everything unless max_checkpoints is set, in which case the oldest on the branch are pruned after each write; the CLI can also prune by count or age. |
| Memory store | LanceDB at <data dir>/memory, table memories; Qdrant Edge shards under memory/qdrant-edge as an option; any backend registered through set_memory_storage_factory. | Memory records: content, scope path, categories, metadata, importance, timestamps, source, a private flag and the embedding vector. | Memory (unified memory), from agents and crews with memory enabled; saves run on a background worker. | None automatic. Records stay until deleted by scope, category, id or age, or until the store is reset (crewai reset-memories). |
| Knowledge store | ChromaDB persistent client in the per-project data directory (the default RAG config); Qdrant is the alternative RAG client. | Chunks of the documents given to knowledge sources, with their embeddings, one collection per knowledge source name. | Knowledge sources on agents and crews save chunks; agents query it before a task. | None automatic; a collection is dropped by reset. |
| CLI settings file | JSON ~/.config/crewai/settings.json, falling back to the temp directory or the working directory when that is not writable; written atomically with mode 0600. | AMP base URL, active organisation name and UUID, OAuth2 provider settings, and tool repository username and password. | The crewai CLI (its config, org, login and enterprise commands). | Until reset or cleared by the CLI. |
| CLI token store | Fernet-encrypted tokens.enc plus its key secret.key in ~/.local/share/crewai/credentials (platform equivalent on macOS and Windows), directory 0700, files 0600. | The platform access token and its expiry. | The CLI after OAuth2 device login; read by every PlusAPI call. | Ignored once expired; deleted on logout or settings reset. |
| Training pickles | Pickle files in the current working directory: training_data.pkl and the trained agents file (default trained_agents_data.pkl, name settable from the CLI). | Per-agent training iterations: initial outputs, human feedback and improved outputs, and the distilled suggestions agents load at kickoff. | crewai train and the crew's train method; read by agents when a trained agents file exists. | The training file is reinitialised at the start of each training run; the trained agents file is kept. |
| File upload cache | aiocache in process memory by default (Redis optional), namespace crewai_uploads, 24 h TTL, at most 1000 entries. | Which file (by content hash) has already been uploaded to which LLM provider, with the provider's file id, URI and expiry. | crewai-files uploaders before sending a file to a provider. | TTL or the provider's own expiry; oldest entries evicted past the cap; lost when the process exits unless Redis is used. |
Each row is a key or index and the access pattern it serves. The JSON checkpoint store, the settings and token files, the training pickles and the upload cache have no index: the file name or cache key is the only lookup, as the last two rows show for the two that encode one.
| Store | Key or index | Kind | Access pattern it serves |
|---|---|---|---|
| Flow state store | flow_states.id | INTEGER primary key, autoincrement | Orders snapshots so the latest for a flow is ORDER BY id DESC; fetches one historical snapshot by row id together with the flow UUID. |
| Flow state store | idx_flow_states_uuid | Index on flow_states(flow_uuid) | Finds a flow's snapshots by flow UUID when restoring or forking. |
| Flow state store | pending_feedback.id | INTEGER primary key, autoincrement | Row identity only; no query reads by it. |
| Flow state store | pending_feedback.flow_uuid | UNIQUE constraint | Keeps at most one pending feedback record per flow, so INSERT OR REPLACE overwrites a previous pause. |
| Flow state store | idx_pending_feedback_uuid | Index on the flow_uuid column of the pending_feedback table | Looks up and deletes the pending record by flow UUID on resume. It covers the same column as the index SQLite builds for the UNIQUE constraint. |
| Kickoff task outputs store | latest_kickoff_task_outputs.task_id | TEXT primary key | One row per task, replaced on re-run. Updates match on task_index and reads order by it, which has no index; the table holds one crew's tasks. |
| Checkpoint store (SQLite) | checkpoints.id | TEXT primary key, <timestamp>_<uuid8> | Loads one checkpoint addressed as db_path#id. The branch and parent_id fields are plain columns: pruning by branch and listing scan the table. |
| Memory store | memories.scope | LanceDB BTREE scalar index | Filters vector search, listing, counting and deletion by scope prefix. There is no approximate vector index: similarity search is a full scan of the table. |
| Knowledge store | knowledge_<collection_name> collections | No declared index | Similarity search per collection; CrewAI declares nothing and relies on ChromaDB's internal vector index. |
| Checkpoint store (JSON) | <branch>/<ts>_<uuid8>_p-<parent>.json | File name | Lineage and ordering are read from the file name and modification time; pruning sorts a branch directory by mtime. |
| File upload cache | upload:<provider>:<file hash> | Cache key | Skips re-uploading a file whose content hash has already been uploaded to the same provider. |
The evidence found eight files with a table definition or storage client; five of them are tests or a tool writing to the user's own database (the Oracle vector search tool). The table adds the source files that hold the other stores, and the two CLI modules that read the SQLite files with their own SQL so the CLI does not have to import the crewai framework. The functions named are described one by one in the table after it.
| File | Store | Functions that touch it | What it does |
|---|---|---|---|
| lib/crewai/src/crewai/flow/persistence/sqlite.py | Flow state store | init_db, save_state, load_state, load_state_from_history, save_pending_feedback, load_pending_feedback, clear_pending_feedback | Implements FlowPersistence on SQLite: creates the schema with retry while locked, appends state snapshots and keeps the pending feedback record. |
| lib/crewai/src/crewai/memory/storage/kickoff_task_outputs_storage.py | Kickoff task outputs store | _initialize_db, add, update, load, delete_all | Stores the latest kickoff's task outputs for replay, adding a missing column on open. |
| lib/crewai/src/crewai/state/provider/sqlite_provider.py | Checkpoint store (SQLite) | checkpoint, acheckpoint, prune, from_checkpoint | The SQLite BaseProvider: inserts, prunes and reads checkpoint rows, sync and async. |
| lib/crewai/src/crewai/state/provider/json_provider.py | Checkpoint store (JSON) | checkpoint, prune, from_checkpoint | The default BaseProvider: one JSON file per checkpoint under a branch directory. |
| lib/crewai/src/crewai/state/provider/utils.py | Checkpoint store | detect_provider | Chooses the provider for a checkpoint being restored. |
| lib/crewai/src/crewai/memory/storage/lancedb_storage.py | Memory store | save, search, delete, reset, _ensure_scope_index, _compact_async | The default StorageBackend on LanceDB, with a scope index and background compaction. |
| lib/crewai/src/crewai/knowledge/storage/knowledge_storage.py | Knowledge store | save, search, reset | Writes and queries knowledge chunks through the RAG client, ChromaDB by default. |
| lib/crewai-core/src/crewai_core/token_manager.py | CLI token store | _get_or_create_key, save_tokens, get_token, clear_tokens | Keeps the platform access token encrypted on disk with a locally generated key. |
| lib/crewai-core/src/crewai_core/settings.py | CLI settings file | get_writable_config_path, dump | Finds a writable settings location and saves the CLI settings model to it. |
| lib/crewai-core/src/crewai_core/user_data.py | Trace consent file (not counted as a store) | update_user_data | Read-modify-writes .crewai_user.json under a file lock. |
| lib/crewai/src/crewai/utilities/file_handler.py | Training pickles | save, load (on PickleHandler) | Pickles and unpickles training data in the working directory under a file lock. |
| lib/crewai-files/src/crewai_files/cache/upload_cache.py | File upload cache | aset, aget, aclear_expired | Tracks provider uploads by content hash in an aiocache backend with TTL and a size cap. |
| lib/cli/src/crewai_cli/task_outputs.py | Kickoff task outputs store | load_task_outputs | Reads the task outputs database directly for the CLI. |
| lib/cli/src/crewai_cli/checkpoint_cli.py | Checkpoint store | _list_sqlite, _prune_sqlite | Lists and prunes SQLite checkpoints with its own copy of the queries. |
| lib/crewai-tools/src/crewai_tools/tools/oracle_vector_search_tool/vector_search.py | User's Oracle database (not a CrewAI store) | create_table, add_texts | Creates a vector table in the user's Oracle database and inserts embedded texts into it. |
| lib/crewai/tests/test_checkpoint_cli.py | Checkpoint store (test fixture) | _create_sqlite_checkpoint | Builds a SQLite checkpoint database for CLI tests, repeating the production schema. |
| lib/crewai-tools/tests/tools/singlestore_search_tool_test.py | Test SingleStore database | sample_table_setup | Creates sample tables in a containerised SingleStore for the search tool tests. |
| lib/crewai-tools/tests/tools/test_nl2sql_security.py | Throwaway SQLite databases | test_write_statements_blocked_by_default, test_dml_actually_persists, test_commented_write_is_committed_when_dml_is_allowed | Exercises the NL2SQL read-only guard against real SQLite files. |
| lib/crewai-files/tests/test_bedrock_async_client.py | Stubbed S3 (test double) | upload_fileobj, delete_object (on _StubS3Client) | Stands in for S3 so the Bedrock uploader can be tested without AWS. |
| Function | Where | What it does |
|---|---|---|
init_db | lib/crewai/src/crewai/flow/persistence/sqlite.py:91 | Creates the flow state tables and indexes, retrying for up to 30 s while another process holds the database locked. |
save_state | lib/crewai/src/crewai/flow/persistence/sqlite.py:193 | Appends a JSON snapshot of the flow's state for a method under the store lock. |
load_state | lib/crewai/src/crewai/flow/persistence/sqlite.py:217 | Returns the most recent saved state for a flow UUID, or None. |
load_state_from_history | lib/crewai/src/crewai/flow/persistence/sqlite.py:249 | Returns one historical snapshot of a flow by its row id, or None when the flow has no such row. |
save_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:285 | In one transaction, appends a state snapshot and replaces the flow's pending feedback record with the pause context and state. |
load_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:329 | Returns the saved state and pause context of a flow waiting for feedback, or None. |
clear_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:368 | Deletes a flow's pending feedback record once it has resumed. |
default_flow_persistence | lib/crewai/src/crewai/flow/persistence/factory.py:48 | Returns the persistence backend from the registered factory, or a new SQLite flow persistence when none is registered. |
set_flow_persistence_factory | lib/crewai/src/crewai/flow/persistence/factory.py:31 | Registers a process-wide factory that supplies the default flow persistence backend. |
_initialize_db | lib/crewai/src/crewai/memory/storage/kickoff_task_outputs_storage.py:32 | Creates the task outputs table in WAL mode and adds the task_key column to an older table that lacks it. |
add | lib/crewai/src/crewai/memory/storage/kickoff_task_outputs_storage.py:78 | Inserts or replaces one task's output, inputs and replay flag. |
update | lib/crewai/src/crewai/memory/storage/kickoff_task_outputs_storage.py:126 | Updates the given columns of the row with a task index. |
load | lib/crewai/src/crewai/memory/storage/kickoff_task_outputs_storage.py:175 | Returns every stored task output ordered by task index. |
delete_all | lib/crewai/src/crewai/memory/storage/kickoff_task_outputs_storage.py:217 | Empties the task outputs table. |
load_task_outputs | lib/cli/src/crewai_cli/task_outputs.py:22 | Opens the task outputs database directly and returns its rows for the CLI to print. |
checkpoint | lib/crewai/src/crewai/state/provider/sqlite_provider.py:59 | Creates the checkpoints table if needed and inserts one checkpoint row, returning db_path#id. |
acheckpoint | lib/crewai/src/crewai/state/provider/sqlite_provider.py:87 | The async version of the SQLite checkpoint write, using aiosqlite. |
prune | lib/crewai/src/crewai/state/provider/sqlite_provider.py:115 | Deletes the oldest checkpoint rows on a branch beyond the number to keep. |
from_checkpoint | lib/crewai/src/crewai/state/provider/sqlite_provider.py:127 | Reads the JSON of the checkpoint addressed by db_path#id. |
checkpoint | lib/crewai/src/crewai/state/provider/json_provider.py:42 | Writes the checkpoint JSON to a new timestamped file under the branch directory and returns its path. |
prune | lib/crewai/src/crewai/state/provider/json_provider.py:98 | Deletes the oldest JSON checkpoint files on a branch, by modification time, beyond the number to keep. |
from_checkpoint | lib/crewai/src/crewai/state/provider/json_provider.py:123 | Reads a JSON checkpoint file and returns its contents. |
detect_provider | lib/crewai/src/crewai/state/provider/utils.py:11 | Returns the SQLite provider when the file (before any #) starts with the SQLite magic bytes, and the JSON provider otherwise. |
_list_sqlite | lib/cli/src/crewai_cli/checkpoint_cli.py:291 | Lists the checkpoints in a SQLite checkpoint database for the CLI. |
_prune_sqlite | lib/cli/src/crewai_cli/checkpoint_cli.py:689 | Deletes SQLite checkpoints beyond a count or older than an age, for crewai checkpoint pruning. |
save | lib/crewai/src/crewai/memory/storage/lancedb_storage.py:314 | Adds memory records to the LanceDB table, creating it on first save, and starts a background compaction every 100 saves. |
search | lib/crewai/src/crewai/memory/storage/lancedb_storage.py:395 | Runs a vector similarity search, optionally filtered by scope prefix, then filters by category and metadata and converts distance to a score. |
delete | lib/crewai/src/crewai/memory/storage/lancedb_storage.py:433 | Deletes memory records by scope, category, id, age or metadata under the store lock, returning how many went. |
reset | lib/crewai/src/crewai/memory/storage/lancedb_storage.py:649 | Drops the whole table, or deletes the records under a scope prefix. |
_ensure_scope_index | lib/crewai/src/crewai/memory/storage/lancedb_storage.py:208 | Creates the BTREE scalar index on the scope column if it is not already there. |
_compact_async | lib/crewai/src/crewai/memory/storage/lancedb_storage.py:238 | Starts a daemon thread that runs LanceDB's optimise to merge fragment files, ignoring failures. |
set_memory_storage_factory | lib/crewai/src/crewai/memory/storage/factory.py:33 | Registers a process-wide factory that turns a storage name into a custom memory backend. |
resolve_memory_storage | lib/crewai/src/crewai/memory/storage/factory.py:48 | Asks the registered factory for a backend for a storage name, returning None when there is no factory or it declines. |
save | lib/crewai/src/crewai/knowledge/storage/knowledge_storage.py:105 | Gets or creates the knowledge collection and adds the document chunks to it. |
search | lib/crewai/src/crewai/knowledge/storage/knowledge_storage.py:59 | Queries the knowledge collection for the chunks closest to the query text. |
reset | lib/crewai/src/crewai/knowledge/storage/knowledge_storage.py:91 | Deletes the knowledge collection. |
_get_or_create_key | lib/crewai-core/src/crewai_core/token_manager.py:32 | Reads the Fernet key from secret.key, or generates one and creates the file exclusively with mode 0600. |
save_tokens | lib/crewai-core/src/crewai_core/token_manager.py:50 | Encrypts the access token and its expiry and writes them atomically to tokens.enc. |
get_token | lib/crewai-core/src/crewai_core/token_manager.py:65 | Decrypts the stored token and returns it, or None when it is missing or expired. |
clear_tokens | lib/crewai-core/src/crewai_core/token_manager.py:80 | Deletes the token file. |
get_writable_config_path | lib/crewai-core/src/crewai_core/settings.py:76 | Returns the first writable settings location among the home config directory, the temp directory and the working directory. |
dump | lib/crewai-core/src/crewai_core/settings.py:246 | Merges the settings that were set into the existing file and writes it atomically with mode 0600, ignoring any error. |
update_user_data | lib/crewai-core/src/crewai_core/user_data.py:54 | Merges updates into .crewai_user.json under a file lock. |
save | lib/crewai/src/crewai/utilities/file_handler.py:144 | Pickles data to the handler's file in the working directory under a file lock. |
load | lib/crewai/src/crewai/utilities/file_handler.py:155 | Unpickles the handler's file, returning an empty dictionary when it is missing or empty. |
aset | lib/crewai-files/src/crewai_files/cache/upload_cache.py:232 | Records that a file was uploaded to a provider, keyed by content hash, with the TTL or the provider's expiry. |
aget | lib/crewai-files/src/crewai_files/cache/upload_cache.py:192 | Returns the cached upload for a file and provider, or None when absent or expired. |
aclear_expired | lib/crewai-files/src/crewai_files/cache/upload_cache.py:347 | Removes every cached upload whose expiry has passed and returns how many it removed. |
lock | lib/crewai-core/src/crewai_core/lock_store.py:80 | Holds a named lock for the duration of a block: a custom backend if one is set, a Redis lock when REDIS_URL is set, otherwise a file lock in the temp directory, with a 120 s default timeout. |
db_storage_path | lib/crewai-core/src/crewai_core/paths.py:16 | Returns, and creates if needed, the per-project data directory under appdirs.user_data_dir. |
create_table | lib/crewai-tools/src/crewai_tools/tools/oracle_vector_search_tool/vector_search.py:693 | Creates the tool's vector table (id, text, JSON metadata, embedding) in the user's Oracle database if it does not exist. |
add_texts | lib/crewai-tools/src/crewai_tools/tools/oracle_vector_search_tool/vector_search.py:805 | Embeds texts and inserts them with their metadata into the Oracle vector table, returning their ids. |
_create_sqlite_checkpoint | lib/crewai/tests/test_checkpoint_cli.py:94 | Test helper that creates a checkpoints table and inserts one checkpoint row. |
sample_table_setup | lib/crewai-tools/tests/tools/singlestore_search_tool_test.py:45 | Test fixture that creates and fills employee and department tables in the test SingleStore database. |
test_write_statements_blocked_by_default | lib/crewai-tools/tests/tools/test_nl2sql_security.py:60 | Checks that each write or DDL statement, including CREATE TABLE, is refused in read-only mode. |
test_dml_actually_persists | lib/crewai-tools/tests/tools/test_nl2sql_security.py:97 | Checks against a file-based SQLite database that an insert is committed when DML is allowed. |
test_commented_write_is_committed_when_dml_is_allowed | lib/crewai-tools/tests/tools/test_nl2sql_security.py:667 | Checks that a delete written with an inline comment is still committed when DML is allowed. |
upload_fileobj | lib/crewai-files/tests/test_bedrock_async_client.py:22 | Test double that records an S3 upload instead of performing it. |
delete_object | lib/crewai-files/tests/test_bedrock_async_client.py:27 | Test double that records an S3 delete instead of performing it. |
No store has a column designed for personal data: there are no user, email or account tables. The stores do, however, hold whatever the developer's users and agents put into a run, in plain text on local disk with no encryption: prompts and task inputs and outputs (task outputs store, checkpoints, flow state snapshots and pending feedback context), remembered facts extracted from conversations (memory store), the documents given to knowledge sources (knowledge store), and human feedback typed during training (training pickles). Any of these can contain personal data, and none has a retention limit. The CLI keeps the developer's own platform access token (encrypted, with the key on the same disk), and the settings file keeps the active organisation and, when set, the tool repository username and password in plain JSON with mode 0600. Nothing personal is stored remotely by these stores; what leaves the machine as telemetry and traces is covered in sections 12 and 16.
Data access sits behind its own layer. Callers in Crew, Flow, Memory and the knowledge classes never issue SQL or open store files themselves; they go through four abstractions, each with a replaceable backend:
| Layer | Interface | Default backend | How it is replaced |
|---|---|---|---|
| Flow persistence | FlowPersistence (abstract base) | SQLite flow persistence | Pass persistence= to a flow or @persist, or register a factory with set_flow_persistence_factory. |
| Memory storage | StorageBackend (protocol) | LanceDB | storage="qdrant-edge", a path, an instance, or a name resolved by set_memory_storage_factory. |
| Checkpoints | BaseProvider (abstract base) | JSON files | provider=SqliteProvider() on CheckpointConfig; restore picks the provider with detect_provider. |
| Knowledge and RAG | KnowledgeStorage over the RAG BaseClient | ChromaDB | The RAG config (Qdrant is the shipped alternative) or a custom embedder per knowledge source. |
Task outputs sit behind TaskOutputStorageHandler without a replaceable backend. Three places bypass the layers: the CLI token and settings files are read and written by their own classes in crewai-core, the training pickles go through PickleHandler, and the CLI's task output and checkpoint commands open the SQLite files directly with their own copy of the queries. Schema changes are handled in code at open time (CREATE TABLE IF NOT EXISTS, and one ALTER TABLE when the task_key column is missing); there is no migration tool.
There is one storage tier per store: a local primary on disk. The only cache tier is the in-memory file upload cache (Redis when configured), plus the in-process tool result cache described in section 15. There is no archive or cold tier, no backup copy, and no search or analytics copy; the memory and knowledge stores are vector search stores in their own right, not copies of another store. The execution traces sent to CrewAI AMP are an outbound event stream rather than a replica of any store here.
Fair. Data access sits behind storage abstractions (StorageBackend, BaseProvider, FlowPersistence, KnowledgeStorage) with pluggable backends; SQLite runs in WAL with busy timeouts and indexes on the lookup keys, LanceDB has a scope index and compaction, and checkpoints prune by branch. Schema evolution is ad hoc (an ALTER TABLE when a column is missing), flow state and memory have no retention or lifecycle, user content is stored unencrypted on local disk, and training data is pickled. The technical assessment gives the reasoning per criterion.
CrewAI is a framework whose job is to call other services: LLM providers to think, tool APIs to act, and optionally the CrewAI AMP platform to deploy, trace and share. It runs no queue, scheduler or object store of its own, so its integrations are outbound HTTP calls made from the developer's process, almost all through a vendor SDK or httpx/requests. Four groups matter at run time.
api.openai.com, api.anthropic.com, generativelanguage.googleapis.com, Bedrock through boto3), and the openai_compatible package names base URLs for OpenRouter, DeepSeek, Cerebras, DeepInfra, DashScope, Ollama and vLLM. When the configured provider is down, every agent step fails: the only retry is the rate-limit retry (3 attempts, 1 s initial delay, 8 s cap, 20% jitter, only for 429/throttling errors), and there is no failover to a second provider.crewai-tools ships 84 tool directories, most wrapping one third-party API (Serper, Brave, Exa, Tavily, Firecrawl, Bright Data, Oxylabs, Browserbase, Apify, Zapier and others). When one is down, the tool returns an error to the agent, which may try another tool or give up within its iteration cap; nothing else breaks. The URL-fetching tools and RAG loaders go through an SSRF-validating adapter that checks the connected peer IP; vendor API tools call their fixed hosts directly (section 8).app.crewai.com (overridable by CREWAI_PLUS_URL) serves CLI login, deploy, the tool and skill repository, platform integration tools and trace batches. When it is down, those commands and tools fail; local crews and flows keep running and trace upload is lost.OTEL_SDK_DISABLED, CREWAI_DISABLE_TELEMETRY or CREWAI_DISABLE_TRACKING is set. Export errors are logged and dropped, so an outage here never stops a run.The evidence counted 446 distinct hosts and SDK clients, and the inventory below has one row for each. Most are not integrations at all. The kind column says what each is, from the evidence's own outbound classification and a read of where each name appears. These kinds group the 446-item list; they are not the evidence's outbound split used under Outbound calls below (76 called from code, 270 named only in tests, 10 local or private, 4 only in comments), which counts a smaller set of hosts. Only the 76 carries over unchanged, so the other counts differ between the two lists.
example.com-style values, made-up test endpoints and hostile-looking hosts used to test the guards.| Host or client | Kind | What breaks when it is down |
|---|---|---|
| 10.0.0.1 | Local or private | Nothing: a private address used as a test value for the SSRF guard and runtime detection. |
| 10.0.0.5 | Local or private | Nothing: a test value for OLLAMA_HOST in the model catalog tests. |
| 169.254.169.254 | Local or private | Nothing: the cloud metadata address, named as a target the guards must refuse. |
| 172.16.0.1 | Local or private | Nothing: a private address used as a test value. |
| 192.168.1.1 | Local or private | Nothing: a private address used as a test value. |
| 192.168.1.5 | Local or private | Nothing: a private address used as a test value. |
| 2p66nmmycsj3.statuspage.io | Test fixture | Nothing: recorded test data; tests run with the network blocked. |
| a.nel.cloudflare.com | Test fixture | Nothing: recorded test data. |
| a2a-protocol.org | Called from code | Nothing: a specification link in A2A extension docstrings, never contacted. |
| a2ui.org | Called from code | Nothing: A2UI extension URIs and JSON schema ids are identifiers, never fetched. |
| about.meta.com | Test fixture | Nothing: recorded test data. |
| accounts.google.com | Test fixture | Nothing: recorded test data. |
| acct.snowflakecomputing.com | Placeholder or example | Nothing: a made-up Snowflake account in tests. |
| actions.zapier.com | Called from code | The Zapier adapter cannot list or run Zapier AI actions. |
| agent.example.com | Placeholder or example | Nothing: an example A2A agent URL. |
| agentops.ai | Test fixture | Nothing: recorded test data. |
| ah-api.merge.dev | Called from code | The Merge Agent Handler tool cannot list or call its tools. |
| ah.merge.dev | Documentation link | Nothing at run time: a link in the tool's README. |
| ai.dev | Called from code | Nothing: prompt text telling the user where to get a Gemini key. |
| ai.stanford.edu | Test fixture | Nothing: recorded test data. |
| aiplatform.googleapis.com | Test fixture | Nothing: recorded Vertex AI test data; at run time the Google SDK reaches Vertex AI without the host being named here. |
| allenai.org | Test fixture | Nothing: recorded test data. |
| amp.example | Placeholder or example | Nothing: a made-up platform URL in tests. |
| amp.exported | Placeholder or example | Nothing: a made-up platform URL in tests. |
| amp.internal | Placeholder or example | Nothing: a made-up platform URL in tests. |
| amp.localhost | Placeholder or example | Nothing: a made-up platform URL in tests. |
| amp.project.test | Placeholder or example | Nothing: a made-up platform URL in tests. |
| amp.test | Placeholder or example | Nothing: a made-up platform URL in tests. |
| analytics.python.org | Test fixture | Nothing: recorded test data. |
| another-domain.auth0.com | Placeholder or example | Nothing: a made-up Auth0 tenant in login tests. |
| another-domain.okta.com | Placeholder or example | Nothing: a made-up Okta tenant in login tests. |
| anywhere.example | Placeholder or example | Nothing: a test value. |
| api-base.openai.com | Placeholder or example | Nothing: a made-up base URL in tests. |
| api.agentops.ai | Test fixture | Nothing: recorded test data. |
| api.ah.merge.dev | Documentation link | Nothing at run time: a link in the Merge tool README. |
| api.anthropic.com | Called from code | The CLI model picker cannot list Anthropic models (it falls back to a curated list); Anthropic LLM calls through the SDK fail. |
| api.brightdata.com | Called from code | The Bright Data dataset, SERP and unlocker tools fail. |
| api.cerebras.ai | Called from code | Cerebras LLM calls fail and the model picker cannot list Cerebras models. |
| api.contextual.ai | Called from code | The Contextual AI parse, query and rerank tools fail. |
| api.deepinfra.com | Called from code | DeepInfra LLM calls fail. |
| api.deepseek.com | Called from code | DeepSeek LLM calls fail. |
| api.exa.ai | Test fixture | Nothing: recorded test data; the Exa tool reaches Exa through its SDK. |
| api.example.com | Placeholder or example | Nothing: an example URL. |
| api.example.test | Placeholder or example | Nothing: an example URL. |
| api.firecrawl.dev | Test fixture | Nothing in tests; at run time the Firecrawl tools reach it through the Firecrawl SDK and fail if it is down. |
| api.github.com | Called from code | crewai template list and crewai template add cannot list or download templates. |
| api.groq.com | Called from code | The model picker cannot list Groq models and falls back to its curated list. |
| api.hubspot.com | Test fixture | Nothing: recorded test data. |
| api.openai.com | Called from code | OpenAI LLM and embedding calls fail; the model picker falls back to its curated list. |
| api.parallel.ai | Called from code | The Parallel search tool fails. |
| api.patronus.ai | Called from code | The Patronus evaluation tools fail. |
| api.pwnedpasswords.com | Test fixture | Nothing: recorded test data. |
| api.scarf.sh | Test fixture | Nothing: recorded test data. |
| api.search.brave.com | Called from code | The Brave search tools fail. |
| api.serply.io | Called from code | The five Serply tools fail. |
| api.stagehand.browserbase.com | Called from code | The Stagehand tool cannot reach its API server. |
| api.star-history.com | Test fixture | Nothing: recorded test data. |
| api.weather.com | Called from code | Nothing: an example MCP reference in a docstring. |
| api.whatsapp.com | Test fixture | Nothing: recorded test data. |
| api.workos.com | Test fixture | Nothing in tests; WorkOS is the default login provider, reached at run time through the configured login domain. |
| apify.com | Documentation link | Nothing at run time: a link in the Apify tool README. |
| app.agentops.ai | Test fixture | Nothing: recorded test data. |
| app.composio.dev | Documentation link | Nothing at run time: a link in a tool README. |
| app.contextual.ai | Documentation link | Nothing at run time: a link in a tool README. |
| app.crewai.com | Called from code | CLI login, deploy, tool and skill publishing, platform tools and trace upload fail; local runs continue. |
| app.hubspot.com | Test fixture | Nothing: recorded test data. |
| app.hyperbrowser.ai | Called from code | Nothing: a sign-up link in a docstring. |
| arxiv.org | Test fixture | Nothing: recorded test data. |
| astral.sh | Documentation link | Nothing at run time: the uv install script URL in the README. |
| attacker.example | Placeholder or example | Nothing: a hostile-looking host used to test the guards. |
| attacker.test | Placeholder or example | Nothing: a hostile-looking host used to test the guards. |
| auth.company.com | Placeholder or example | Nothing: a made-up identity provider in tests. |
| auth.enterprise.com | Placeholder or example | Nothing: a made-up identity provider in tests. |
| auth.example.com | Placeholder or example | Nothing: a made-up identity provider in tests. |
| avatars.githubusercontent.com | Test fixture | Nothing: recorded test data. |
| base-url.openai.com | Placeholder or example | Nothing: a made-up base URL in tests. |
| bat.bing.com | Test fixture | Nothing: recorded test data. |
| bedrock-runtime.us-east-1.amazonaws.com | Test fixture | Nothing in tests; at run time boto3 reaches the regional Bedrock endpoint without the host being named here. |
| bedrock-runtime.us-west-2.amazonaws.com | Test fixture | Nothing: recorded Bedrock test data. |
| bedrock-runtime.vcr.amazonaws.com | Placeholder or example | Nothing: conftest.py rewrites recorded Bedrock hosts to this placeholder. |
| blog.apify.com | Documentation link | Nothing at run time: a README link. |
| blog.crewai.com | Documentation link | Nothing at run time: a README link. |
| blogs.microsoft.com | Test fixture | Nothing: recorded test data. |
| boto3 | SDK client | The AWS SDK: Bedrock LLM calls, the Bedrock knowledge base and agent tools, the S3 reader and writer tools and Bedrock file uploads fail when AWS is unreachable. |
| brightdata.com | Documentation link | Nothing at run time: a README link. |
| browserbase.com | Documentation link | Nothing at run time: a README link. |
| bucket.example | Placeholder or example | Nothing: a made-up bucket in tests. |
| bucket.r2.cloudflarestorage.com | Called from code | Nothing: a presigned-URL example in a docstring. |
| camo.githubusercontent.com | Test fixture | Nothing: recorded test data. |
| cdn.example.com | Placeholder or example | Nothing: an example URL. |
| cdn.jsdelivr.net | Test fixture | Nothing: recorded test data. |
| cdn.theconversation.com | Test fixture | Nothing: recorded test data. |
| cdnjs.cloudflare.com | Called from code | The flow visualisation page cannot export to PNG or PDF (html2canvas and jspdf do not load); the graph still renders. |
| cerebras.ai | Called from code | Nothing: a link in a docstring's provider list. |
| chatg.pt | Documentation link | Nothing at run time: a link in a project template README. |
| checkout.stripe.com | Test fixture | Nothing: recorded test data. |
| cli.github.com | Called from code | Nothing: an install hint printed by the release tool. |
| cloud.google.com | Test fixture | Nothing: recorded test data and one docs link in a comment. |
| cognitiveservices.azure.com | Called from code | Nothing directly: it is an OAuth scope identifier; the token request goes to Entra ID. |
| collector.example | Placeholder or example | Nothing: a made-up telemetry collector in tests. |
| collector.github.com | Test fixture | Nothing: recorded test data. |
| collector.invalid | Placeholder or example | Nothing: an unresolvable test value. |
| community.crewai.com | Documentation link | Nothing at run time: a README link. |
| configured-url.com | Placeholder or example | Nothing: a test value. |
| configured.example.test | Placeholder or example | Nothing: a test value. |
| connect.useparagon.com | Test fixture | Nothing: recorded test data. |
| context7.com | Called from code | Nothing: context7.json registers the repository with the Context7 docs index; no code calls it. |
| corp.example | Placeholder or example | Nothing: a test value. |
| cp-attacker.example | Placeholder or example | Nothing: a hostile-looking host used to test the guards. |
| crewai-production.s3.us-west-2.amazonaws.com | Documentation link | Nothing in this repository: named in devtools' AGENTS.md evaluation items as the bucket platform skill downloads come from. |
| crewai.com | Called from code | Nothing: the HTTP-Referer value sent to OpenRouter and the package homepage. |
| custom-api.com | Placeholder or example | Nothing: a test value. |
| custom-url-from-env.com | Placeholder or example | Nothing: a test value. |
| custom-url.com | Placeholder or example | Nothing: a test value. |
| custom.crewai.com | Placeholder or example | Nothing: a test value. |
| custom.deepseek.com | Placeholder or example | Nothing: a test value. |
| custom.exa.api.com | Placeholder or example | Nothing: a test value. |
| custom.example.com | Placeholder or example | Nothing: a test value. |
| custom.openai.com | Placeholder or example | Nothing: a test value. |
| custom.openrouter.ai | Placeholder or example | Nothing: a test value. |
| dashboard.exa.ai | Documentation link | Nothing at run time: a README link. |
| dashboard.oxylabs.io | Called from code | Nothing: a sign-up link in Oxylabs tool docstrings. |
| dashboard.tools.crewai.com | Test fixture | Nothing: recorded test data. |
| dashscope-intl.aliyuncs.com | Called from code | Alibaba DashScope LLM calls fail. |
| dashscope.aliyun.com | Called from code | Nothing: a link in a docstring's provider list. |
| deepinfra.com | Called from code | Nothing: a link in a docstring's provider list. |
| deepmind.google | Test fixture | Nothing: recorded test data. |
| deepseek.com | Called from code | Nothing: a link in a docstring's provider list. |
| descriptusercontent.com | Test fixture | Nothing: recorded test data. |
| dev.auth0.com | Placeholder or example | Nothing: a made-up Auth0 tenant in tests. |
| dev.localhost | Placeholder or example | Nothing: a test value. |
| dev.okta.com | Placeholder or example | Nothing: a made-up Okta tenant in tests. |
| developers.facebook.com | Documentation link | Nothing at run time: a reference URL in the platform app catalog. |
| developers.google.com | Called from code | Nothing: reference links in Serply docstrings and scope strings in the platform catalog. |
| developers.oxylabs.io | Called from code | Nothing: documentation links in Oxylabs tool docstrings. |
| dify.ai | Test fixture | Nothing: recorded test data. |
| discord.com | Documentation link | Nothing at run time: a link in project template READMEs. |
| discord.gg | Test fixture | Nothing: recorded test data. |
| docs-platform.crewai.com | Called from code | Nothing: a link in a legacy-token warning. |
| docs.ag2.ai | Test fixture | Nothing: recorded test data. |
| docs.agentops.ai | Test fixture | Nothing: recorded test data. |
| docs.ah.merge.dev | Documentation link | Nothing at run time: a README link. |
| docs.anthropic.com | Test fixture | Nothing: recorded test data. |
| docs.apify.com | Called from code | Nothing: documentation links in the Apify tool. |
| docs.astral.sh | Called from code | Nothing: an install hint in deploy validation. |
| docs.aws.amazon.com | Documentation link | Nothing at run time: README and docstring links. |
| docs.browserbase.com | Documentation link | Nothing at run time: README links. |
| docs.camel-ai.org | Test fixture | Nothing: recorded test data. |
| docs.cloud.google.com | Called from code | Nothing: a migration link in Vertex AI deprecation warnings. |
| docs.cohere.com | Test fixture | Nothing: recorded test data. |
| docs.contextual.ai | Documentation link | Nothing at run time: README links. |
| docs.couchbase.com | Documentation link | Nothing at run time: README links. |
| docs.crewai | Documentation link | Nothing: a truncated fragment of docs.crewai.com picked up by the scan. |
| docs.crewai.com | Called from code | Nothing: links in error, help and package metadata. |
| docs.example.com | Placeholder or example | Nothing: an example URL. |
| docs.firecrawl.dev | Called from code | Nothing: a docstring link in scripts/wharf_runner.py. |
| docs.getcargo.io | Test fixture | Nothing: recorded test data. |
| docs.github.com | Documentation link | Nothing at run time: comments in .github workflow and Dependabot files. |
| docs.google.com | Test fixture | Nothing: recorded test data. |
| docs.hyperbrowser.ai | Called from code | Nothing: documentation links in a field description. |
| docs.langflow.org | Test fixture | Nothing: recorded test data. |
| docs.litellm.ai | Documentation link | Nothing at run time: a comment link. |
| docs.llamaindex.ai | Documentation link | Nothing at run time: a README link. |
| docs.mdb.ai | Documentation link | Nothing at run time: a README link. |
| docs.mistral.ai | Test fixture | Nothing: recorded test data. |
| docs.multion.ai | Documentation link | Nothing at run time: a README link. |
| docs.parallel.ai | Documentation link | Nothing at run time: README links. |
| docs.praison.ai | Test fixture | Nothing: recorded test data. |
| docs.stagehand.dev | Documentation link | Nothing at run time: README links. |
| docs.swarmzero.ai | Test fixture | Nothing: recorded test data. |
| docs.vectorize.io | Test fixture | Nothing: recorded test data. |
| doi.org | Test fixture | Nothing: recorded test data. |
| download.pytorch.org | Called from code | Installing with uv cannot resolve torch from the pinned PyTorch index; nothing at run time. |
| drive.google.com | Test fixture | Nothing: recorded test data. |
| e2b.dev | Documentation link | Nothing at run time: a README link. |
| en.wikipedia.org | Test fixture | Nothing: recorded test data. |
| entelligence.ai | Test fixture | Nothing: recorded test data. |
| enterprise.example.com | Placeholder or example | Nothing: a test value. |
| env.deepseek.com | Placeholder or example | Nothing: a test value. |
| env.exa.api.com | Placeholder or example | Nothing: a test value. |
| env.openai.com | Placeholder or example | Nothing: a test value. |
| ethicsinaiai.ieee.org | Test fixture | Nothing: recorded test data. |
| evil.example | Placeholder or example | Nothing: a hostile-looking host used to test the guards. |
| evil.example.com | Placeholder or example | Nothing: a hostile-looking host used to test the guards. |
| evil.test | Placeholder or example | Nothing: a hostile-looking host used to test the guards. |
| evolve.crewai.test | Placeholder or example | Nothing: a test value. |
| exa.ai | Documentation link | Nothing at run time: a README link. |
| example-spa.com | Placeholder or example | Nothing: an example URL. |
| example.azure.com | Placeholder or example | Nothing: an example URL. |
| example.com | Placeholder or example | Nothing: an example URL. |
| example.invalid | Placeholder or example | Nothing: an unresolvable test value. |
| example.openai.azure.com | Placeholder or example | Nothing: an example URL. |
| example.test | Placeholder or example | Nothing: a test value. |
| example.weaviate.cloud | Placeholder or example | Nothing: an example URL. |
| explicit-url.com | Placeholder or example | Nothing: a test value. |
| explicit.deepseek.com | Placeholder or example | Nothing: a test value. |
| export.arxiv.org | Called from code | The arXiv paper tool fails. It queries over plain http. |
| external.mcp.com | Placeholder or example | Nothing: a made-up MCP server in tests. |
| facebook.com | Test fixture | Nothing: recorded test data. |
| factory.example | Placeholder or example | Nothing: a test value. |
| fake-azure-endpoint.openai.azure.com | Test fixture | Nothing: a fake endpoint in recorded Azure tests. |
| fake.crewai.com | Placeholder or example | Nothing: a test value. |
| fallback-url.com | Placeholder or example | Nothing: a test value. |
| fallback.openai.com | Placeholder or example | Nothing: a test value. |
| fever.ai | Test fixture | Nothing: recorded test data. |
| files.example.com | Placeholder or example | Nothing: an example URL. |
| firecrawl.dev | Test fixture | Nothing: recorded test data. |
| first.example.com | Placeholder or example | Nothing: an example URL. |
| fk4bvu0n5qp.typeform.com | Test fixture | Nothing: recorded test data. |
| flink.apache.org | Test fixture | Nothing: recorded test data. |
| forms.hscollectedforms.net | Test fixture | Nothing: recorded test data. |
| gateway.corp | Placeholder or example | Nothing: a test value. |
| gateway.example | Placeholder or example | Nothing: a test value. |
| generativelanguage.googleapis.com | Called from code | Gemini LLM calls fail; the model picker falls back to its curated list. |
| gh.io | Documentation link | Nothing at run time: comments in codeql.yml. |
| gist.github.com | Test fixture | Nothing: recorded test data. |
| git.io | Test fixture | Nothing: recorded test data. |
| github.com | Called from code | The GitHub search tool and loader cannot fetch repositories; CI links stop resolving. |
| github.com.evil.example | Placeholder or example | Nothing: a lookalike host used to test URL validation. |
| github.githubassets.com | Test fixture | Nothing: recorded test data. |
| gitlab.com | Test fixture | Nothing: recorded test data. |
| gohugo.io | Test fixture | Nothing: recorded test data. |
| google.com | Called from code | Nothing: an example navigation target in Stagehand prompts. |
| google.serper.dev | Called from code | The Serper search tool fails. |
| httpbin.org | Called from code | Nothing: an example target in the Stagehand example script. |
| huggingface.co | Test fixture | Nothing: recorded test data. |
| hyperbrowser.ai | Documentation link | Nothing at run time: a README link. |
| identity.org | Test fixture | Nothing: recorded test data. |
| image1.com | Placeholder or example | Nothing: a test value. |
| images.theconversation.com | Test fixture | Nothing: recorded test data. |
| img.co | Placeholder or example | Nothing: a test value. |
| img.shields.io | Documentation link | Nothing at run time: README badges, also present in recorded test data. |
| img.youtube.com | Documentation link | Nothing at run time: README video thumbnails. |
| infer.roboflow.com | Called from code | The Roboflow embedding provider fails. |
| infotabs.github.io | Test fixture | Nothing: recorded test data. |
| ip.oxylabs.io | Documentation link | Nothing at run time: a README example. |
| js-na1.hs-scripts.com | Test fixture | Nothing: recorded test data. |
| js.hs-analytics.net | Test fixture | Nothing: recorded test data. |
| js.hscollectedforms.net | Test fixture | Nothing: recorded test data. |
| js.langchain.com | Test fixture | Nothing: recorded test data. |
| js.sentry-cdn.com | Test fixture | Nothing: recorded test data. |
| js.usemessages.com | Test fixture | Nothing: recorded test data. |
| keycloak.example.com | Placeholder or example | Nothing: a made-up Keycloak server in login tests. |
| kodexolabs.com | Test fixture | Nothing: recorded test data. |
| learn.crewai.com | Documentation link | Nothing at run time: a README link. |
| learn.microsoft.com | Called from code | Nothing: a docstring link. |
| legacy-env.openai.com | Placeholder or example | Nothing: a test value. |
| lilianweng.github.io | Test fixture | Nothing: recorded test data. |
| litellm.ai | Test fixture | Nothing: recorded test data. |
| llama.meta.com | Test fixture | Nothing: recorded test data. |
| lmsys.org | Test fixture | Nothing: recorded test data. |
| localhost | Local or private | Default address of a local Ollama server (port 11434), vLLM and similar local services; if nothing listens there, those LLM calls and the Ollama model listing fail. |
| localhost.example | Placeholder or example | Nothing: a test value. |
| login.company.com | Placeholder or example | Nothing: a made-up identity provider in tests. |
| login.example.com | Placeholder or example | Nothing: a made-up identity provider in tests. |
| login.microsoftonline.com | Called from code | crewai login fails for organisations that use the Entra ID provider. |
| login.myapp.io | Placeholder or example | Nothing: a made-up identity provider in tests. |
| lower.test | Placeholder or example | Nothing: a test value. |
| machine-proxy.test | Placeholder or example | Nothing: a made-up proxy in proxy-pinning tests. |
| mcp.brightdata.com | Test fixture | Nothing: recorded test data. |
| mcp.exa.ai | Test fixture | Nothing: recorded test data. |
| mcp.example.com | Placeholder or example | Nothing: a made-up MCP server. |
| mcp.example.test | Placeholder or example | Nothing: a made-up MCP server. |
| mcp.firecrawl.dev | Called from code | Only the maintainer script scripts/wharf_runner.py cannot reach the Firecrawl MCP server. |
| mcp.github.com | Test fixture | Nothing: recorded test data. |
| mcp.notion.so | Test fixture | Nothing: recorded test data. |
| mdb.ai | Called from code | The AI Minds tool fails. |
| media.istockphoto.com | Test fixture | Nothing: recorded test data. |
| media.springernature.com | Test fixture | Nothing: recorded test data. |
| medium.com | Test fixture | Nothing: recorded test data. |
| mendable.ai | Test fixture | Nothing: recorded test data. |
| merge.dev | Documentation link | Nothing at run time: a README link. |
| mindsdb.com | Called from code | Nothing: a link in the AI Minds tool description. |
| mintlify.com | Documentation link | Nothing at run time: a CONTRIBUTING link. |
| mintlify.s3-us-west-1.amazonaws.com | Test fixture | Nothing: recorded test data. |
| miro.medium.com | Test fixture | Nothing: recorded test data. |
| models.example | Placeholder or example | Nothing: a test value. |
| models.inference.ai.azure.com | Test fixture | Nothing: recorded test data. |
| msblogs.thesourcemediaassets.com | Test fixture | Nothing: recorded test data. |
| my-company.auth0.com | Placeholder or example | Nothing: a made-up Auth0 tenant. |
| my-company.okta.com | Placeholder or example | Nothing: a made-up Okta tenant. |
| my-dashscope.example.com | Placeholder or example | Nothing: a test value. |
| my-server.com | Placeholder or example | Nothing: an example MCP server. |
| my-vllm.internal | Placeholder or example | Nothing: an example vLLM server. |
| myresource.openai.azure.com | Called from code | Nothing: an example endpoint in a docstring. |
| n.co | Placeholder or example | Nothing: a test value. |
| news.ycombinator.com | Test fixture | Nothing: recorded test data. |
| news1.com | Placeholder or example | Nothing: a test value. |
| not-logged-in.test | Placeholder or example | Nothing: a test value. |
| oauth2.googleapis.com | Test fixture | Nothing: recorded test data. |
| ollama.ai | Called from code | Nothing: a link in a docstring's provider list. |
| ollama.example.com | Placeholder or example | Nothing: a test value. |
| openai.com | Test fixture | Nothing: recorded test data. |
| openrouter.ai | Called from code | OpenRouter LLM and embedding calls fail. |
| opensource.org | Documentation link | Nothing at run time: the README licence link. |
| opentelemetry.io | Called from code | Nothing: specification links in tracing modules. |
| optimize.crewai.test | Placeholder or example | Nothing: a test value. |
| optimize.example | Placeholder or example | Nothing: a test value. |
| org-account.snowflakecomputing.com | Placeholder or example | Nothing: a made-up Snowflake account. |
| ossrdbms-aad.database.windows.net | Called from code | Nothing directly: an Entra ID token scope for the Azure DocumentDB tool; the token request goes to Entra ID. |
| other-private.example | Placeholder or example | Nothing: a test value. |
| other.com | Placeholder or example | Nothing: a test value. |
| other.example | Placeholder or example | Nothing: a test value. |
| other.mcp.com | Placeholder or example | Nothing: a made-up MCP server. |
| oxylabs.io | Documentation link | Nothing at run time: README links. |
| people.eecs.berkeley.edu | Test fixture | Nothing: recorded test data. |
| pepy.tech | Test fixture | Nothing: recorded test data. |
| platform.example.test | Placeholder or example | Nothing: a test value. |
| polyfill.io | Test fixture | Nothing: recorded test data. |
| pre-commit.com | Documentation link | Nothing at run time: a CONTRIBUTING link. |
| priority.openai.com | Placeholder or example | Nothing: a test value. |
| prod.auth0.com | Placeholder or example | Nothing: a made-up Auth0 tenant. |
| prod.okta.com | Placeholder or example | Nothing: a made-up Okta tenant. |
| project-proxy.test | Placeholder or example | Nothing: a made-up proxy in proxy-pinning tests. |
| promptarmor.com | Test fixture | Nothing: recorded test data. |
| proxy.corp | Placeholder or example | Nothing: a made-up proxy in tests. |
| proxy.example | Placeholder or example | Nothing: a made-up proxy in tests. |
| proxy.example.com | Placeholder or example | Nothing: a made-up proxy in tests. |
| public.example | Placeholder or example | Nothing: a test value. |
| px.ads.linkedin.com | Test fixture | Nothing: recorded test data. |
| pypi-camo.freetls.fastly.net | Test fixture | Nothing: recorded test data. |
| pypi.org | Called from code | The console's new-version notice is skipped; releases cannot be published or checked. |
| python.langchain.com | Test fixture | Nothing: recorded test data. |
| qdrant.tech | Documentation link | Nothing at run time: a README link. |
| r.co | Placeholder or example | Nothing: a test value. |
| r.jina.ai | Called from code | The Jina scrape tool fails. |
| r.openai.azure.com | Placeholder or example | Nothing: a test value. |
| raw.githubusercontent.com | Called from code | The CLI model picker cannot refresh the LiteLLM catalog and uses its cache or curated list; the maintainers' catalog refresh fails. |
| realtime.oxylabs.io | Test fixture | Nothing in tests; the Oxylabs tools reach Oxylabs through their SDK. |
| reddit.com | Test fixture | Nothing: recorded test data. |
| repository-images.githubusercontent.com | Test fixture | Nothing: recorded test data. |
| res.openai.azure.com | Placeholder or example | Nothing: a test value. |
| research.ibm.com | Test fixture | Nothing: recorded test data. |
| run.pstmn.io | Test fixture | Nothing: recorded test data. |
| s3 | SDK client | The boto3 S3 client: the S3 reader and writer tools and Bedrock file uploads fail. |
| safe.example | Placeholder or example | Nothing: a test value. |
| schema.org | Test fixture | Nothing: recorded test data. |
| schemas.openxmlformats.org | Test fixture | Nothing: XML namespaces in test documents. |
| scrape.serper.dev | Called from code | The Serper scrape tool fails. |
| scrapegraphai.com | Documentation link | Nothing at run time: a README link. |
| scrapfly.io | Documentation link | Nothing at run time: a README link. |
| second.example.com | Placeholder or example | Nothing: an example URL. |
| secure.test | Placeholder or example | Nothing: a test value. |
| serpapi.com | Called from code | Nothing: a link in the missing-key error message. |
| serper.dev | Documentation link | Nothing at run time: a README link. |
| serply.io | Documentation link | Nothing at run time: a README link. |
| server.com | Called from code | Nothing: an example MCP URL in a field description. |
| share.descript.com | Test fixture | Nothing: recorded test data. |
| shell.amp.test | Placeholder or example | Nothing: a test value. |
| singlestoredb-python.labs.singlest | Documentation link | Nothing at run time: a truncated README link to the SingleStore Python docs. |
| skills.sh | Documentation link | Nothing at run time: a README link. |
| spider.cloud | Documentation link | Nothing at run time: README links. |
| ssl.google-analytics.com | Test fixture | Nothing: recorded test data. |
| sso.company.com | Placeholder or example | Nothing: a made-up identity provider. |
| sso.enterprise.com | Placeholder or example | Nothing: a made-up identity provider. |
| stackoverflow.com | Called from code | Nothing: a reference in a comment. |
| stagehand.dev | Documentation link | Nothing at run time: README links. |
| static.cloudflareinsights.com | Test fixture | Nothing: recorded test data. |
| status.firecrawl.dev | Test fixture | Nothing: recorded test data. |
| storage.googleapis.com | Test fixture | Nothing: recorded test data. |
| studio.crewai.com | Documentation link | Nothing at run time: a README link. |
| swarmzero.ai | Test fixture | Nothing: recorded test data. |
| telegram.me | Test fixture | Nothing: recorded test data. |
| telemetry.crewai.com | Called from code | Telemetry spans are not exported; export errors are logged and dropped, and runs are unaffected. |
| test-ai.services.example.com | Placeholder or example | Nothing: a test value. |
| test-api-base.com | Placeholder or example | Nothing: a test value. |
| test-domain.auth0.com | Placeholder or example | Nothing: a made-up Auth0 tenant in login tests. |
| test-domain.okta.com | Placeholder or example | Nothing: a made-up Okta tenant in login tests. |
| test-endpoint.com | Placeholder or example | Nothing: a test value. |
| test-resource.openai.azure.com | Placeholder or example | Nothing: a test value. |
| test.com | Placeholder or example | Nothing: a test value. |
| test.litellm.proxy | Placeholder or example | Nothing: a test value. |
| test.openai.azure.com | Placeholder or example | Nothing: a fake Azure endpoint in tests. |
| test1.com | Placeholder or example | Nothing: a test value. |
| test1.openai.azure.com | Placeholder or example | Nothing: a test value. |
| test2.com | Placeholder or example | Nothing: a test value. |
| test2.openai.azure.com | Placeholder or example | Nothing: a test value. |
| theconversation.com | Test fixture | Nothing: recorded test data. |
| this-host-definitely-does-not-exist-abc123.com | Placeholder or example | Nothing: an unresolvable host used to test failure handling. |
| trendshift.io | Documentation link | Nothing at run time: a README badge. |
| twitter.com | Documentation link | Nothing at run time: a README link. |
| unpkg.com | Called from code | The flow visualisation page cannot draw its graph: vis-network does not load. |
| unrelated.example | Placeholder or example | Nothing: a test value. |
| upper.test | Placeholder or example | Nothing: a test value. |
| url-from-env.com | Placeholder or example | Nothing: a test value. |
| url-from-settings.com | Placeholder or example | Nothing: a test value. |
| us-central1-aiplatform.googleapis.com | Test fixture | Nothing: recorded Vertex AI test data. |
| us-south.ml.cloud.ibm.com | Test fixture | Nothing: recorded test data. |
| us.i.posthog.com | Test fixture | Nothing: recorded test data. |
| v.co | Placeholder or example | Nothing: a test value. |
| vllm.example | Placeholder or example | Nothing: an example vLLM server. |
| weaviate.io | Documentation link | Nothing at run time: README links. |
| web-scraping.dev | Documentation link | Nothing at run time: a README example target. |
| web.archive.org | Test fixture | Nothing: recorded test data. |
| worker-actionkit.tools.crewai.com | Test fixture | Nothing: recorded test data. |
| www.amazon.com | Documentation link | Nothing at run time: a README example target. |
| www.bing.com | Called from code | Nothing directly: a search URL handed to Bright Data, which fetches it. |
| www.browserbase.com | Documentation link | Nothing at run time: README links. |
| www.camel-ai.org | Test fixture | Nothing: recorded test data. |
| www.conventionalcommits.org | Documentation link | Nothing at run time: a CONTRIBUTING link. |
| www.crewai.com | Documentation link | Nothing at run time: a template README link. |
| www.daytona.io | Documentation link | Nothing at run time: a README link. |
| www.deeplearning.ai | Documentation link | Nothing at run time: README course links. |
| www.example.com | Placeholder or example | Nothing: an example URL. |
| www.firecrawl.dev | Test fixture | Nothing: recorded test data. |
| www.gartner.com | Test fixture | Nothing: recorded test data. |
| www.gatsby.ucl.ac.uk | Test fixture | Nothing: recorded test data. |
| www.goodreads.com | Test fixture | Nothing: recorded test data. |
| www.google-analytics.com | Test fixture | Nothing: recorded test data. |
| www.google.com | Called from code | Nothing directly: a search URL handed to Bright Data and a Referer header value in two scrape tools. |
| www.google.com.br | Test fixture | Nothing: recorded test data. |
| www.googleapis.com | Documentation link | Nothing at run time: OAuth scope strings in the platform app catalog. |
| www.googletagmanager.com | Test fixture | Nothing: recorded test data. |
| www.gstatic.com | Test fixture | Nothing: recorded test data. |
| www.ibm.com | Documentation link | Nothing at run time: a reference URL in source text. |
| www.lesswrong.com | Test fixture | Nothing: recorded test data. |
| www.linkedin.com | Called from code | Nothing: a share link in the first-contributor welcome comment. |
| www.llama.com | Test fixture | Nothing: recorded test data. |
| www.macrotrends.net | Test fixture | Nothing: recorded test data. |
| www.mckinsey.com | Test fixture | Nothing: recorded test data. |
| www.mendable.ai | Test fixture | Nothing: recorded test data. |
| www.mintlify.com | Documentation link | Nothing at run time: a README link. |
| www.mongodb.com | Documentation link | Nothing at run time: README and docstring links. |
| www.nature.com | Documentation link | Nothing at run time: a README example. |
| www.nippon.com | Test fixture | Nothing: recorded test data. |
| www.pabbly.com | Test fixture | Nothing: recorded test data. |
| www.rolustech.com | Test fixture | Nothing: recorded test data. |
| www.scrapfly.io | Documentation link | Nothing at run time: a README link. |
| www.stagehand.dev | Called from code | Nothing: an example target in the Stagehand example script. |
| www.un.org | Test fixture | Nothing: recorded test data. |
| www.us.maguireshoes.com | Test fixture | Nothing: recorded test data. |
| www.w3.org | Documentation link | Nothing at run time: an XML or SVG namespace string. |
| www.youtube.com | Documentation link | Nothing at run time: README video links, also in recorded test data. |
| x.blob.core.windows.net | Placeholder or example | Nothing: a test value. |
| x.com | Called from code | Nothing: a share link in the first-contributor welcome comment. |
| x.openai.azure.com | Placeholder or example | Nothing: a test value. |
| xxx.xxx.xxx.xxx | Placeholder or example | Nothing: a placeholder address. |
| yandex.com | Called from code | Nothing directly: a search URL handed to Bright Data. |
| your-qdrant-cluster-url.com | Placeholder or example | Nothing: a placeholder in an example. |
| your-resource.openai.azure.com | Placeholder or example | Nothing: a placeholder in an example. |
| your-weaviate-cluster-url.com | Placeholder or example | Nothing: a placeholder in an example. |
| your-workspace.cloud.databricks.com | Placeholder or example | Nothing: a placeholder in an example. |
| youtube.com | Documentation link | Nothing at run time: a README link. |
| zapier.com | Documentation link | Nothing at run time: a README link. |
| zeus.tools.crewai.com | Test fixture | Nothing: recorded test data. |
| zeus.useparagon.com | Test fixture | Nothing: recorded test data. |
The evidence lists 76 hosts called from code, 270 named only in tests, 10 local or private and 4 only in comments. It found no install hooks and 43 risk-pattern lines, 40 of them in tests. Reading the 76 shows two kinds. Thirty-five are hosts the code really connects to (or, for the PyTorch index, the installer does): LLM providers, tool APIs, the CrewAI platform, telemetry, PyPI, GitHub, and two CDNs the flow visualisation page loads. The rest are named only in a string that never becomes a request: a docstring, a help or error message, a schema identifier, an OAuth scope, an example prompt. Every connection is one a reader would expect from an LLM agent framework with a tool library and a vendor platform. Nothing calls a paste site, a tunnel, a webhook relay or a raw IP address.
Three outbound details are worth a reader's attention, none of them hostile. The flow visualisation page loads vis-network 9.1.2 from unpkg.com and html2canvas 1.4.1 and jspdf 2.5.1 from cdnjs.cloudflare.com, version-pinned but without subresource integrity. The arXiv tool queries export.arxiv.org over plain http. And CREWAI_FACTORY=true turns off TLS verification for platform integration calls (section 9).
| Host | Where it is called | What for | Expected? |
|---|---|---|---|
| api.openai.com | lib/cli/src/crewai_cli/model_catalog.py:212 | The CLI model picker lists OpenAI models (GET /v1/models) with the user's key; the OpenAI SDK reaches the same host for LLM calls. | Yes: an LLM provider. |
| github.com | lib/crewai-tools/src/crewai_tools/rag/loaders/github_loader.py:29; tools/github_search_tool/github_search_tool.py:60; .github/workflows/ftc-require-issue.yml:134 | The GitHub loader and search tool fetch repositories an agent is pointed at; workflows and package metadata link to the repository. | Yes. |
| app.crewai.com | lib/crewai-core/src/crewai_core/constants.py:14; lib/cli/src/crewai_cli/constants.py:4; crewai_tools/adapters/enterprise_adapter.py:14 | CrewAI AMP platform API: login, deploy, tool and skill repository, platform tools, trace batches. | Yes: the vendor's own platform, reached by commands and features the user invokes and by trace upload when tracing is on. |
| api.anthropic.com | lib/cli/src/crewai_cli/model_catalog.py:243 | The model picker lists Anthropic models. | Yes: an LLM provider. |
| generativelanguage.googleapis.com | lib/cli/src/crewai_cli/model_catalog.py:265 | The model picker lists Gemini models, paging up to ten times. | Yes: an LLM provider. |
| docs.firecrawl.dev | scripts/wharf_runner.py:19 | A docstring link; not contacted. | Yes. |
| www.google.com | crewai_tools/tools/brightdata_tool/brightdata_serp.py:137; scrape_website_tool.py:41 | A Google search URL handed to Bright Data, and a Referer header value; crewai does not connect to it. | Yes. |
| google.com | crewai_tools/tools/stagehand_tool/example.py:54; stagehand_tool.py:113 | An example navigation target in prompts and docstrings. | Yes. |
| x.com | .github/workflows/first-contributor-welcome.yml:72 | A share link in a PR welcome comment; not contacted. | Yes. |
| telemetry.crewai.com | lib/crewai/src/crewai/telemetry/constants.py:9; lib/crewai-core/src/crewai_core/telemetry.py:45 | OTLP export of anonymous telemetry spans to port 4319. | Expected for a vendor framework; on by default and switched off by the three opt-out variables. |
| pypi.org | lib/crewai-core/src/crewai_core/version.py:126; lib/devtools/src/crewai_devtools/cli.py:1742; .github/workflows/nightly.yml:130 | The console's new-version check, the release tool's published-version check, and CI publishing. | Yes. |
| cdnjs.cloudflare.com | lib/crewai/src/crewai/flow/visualization/assets/interactive.js:2282,2299,2303 | html2canvas 1.4.1 and jspdf 2.5.1, loaded by the browser to export the flow graph. | Expected; version-pinned but without subresource integrity. |
| api.github.com | lib/cli/src/crewai_cli/remote_template/main.py:22 | crewai template lists the crewAIInc organisation's template repositories and downloads a zipball. | Yes. |
| docs.crewai.com | lib/cli/src/crewai_cli/update_crew.py:18; lib/crewai/src/crewai/llm.py:361 | Links in help and error messages and package metadata. | Yes. |
| login.microsoftonline.com | lib/crewai-core/src/crewai_core/auth/providers/entra_id.py:49 | OAuth2 device flow for organisations that log in with Entra ID. | Yes. |
| openrouter.ai | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:49; rag/embeddings/providers/openrouter/openrouter_provider.py:42 | OpenRouter LLM and embedding base URL. | Yes: an LLM provider. |
| api.search.brave.com | crewai_tools/tools/brave_search_tool/brave_image_tool.py:24 (8 lines) | Brave web, image, news, local and LLM-context search. | Yes: a tool API. |
| www.linkedin.com | .github/workflows/first-contributor-welcome.yml:73 | A share link in a PR welcome comment; not contacted. | Yes. |
| crewai.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:52 | The HTTP-Referer value sent to OpenRouter, and package homepages. | Yes; not contacted. |
| myresource.openai.azure.com | lib/crewai/src/crewai/llms/providers/azure/completion.py:288 | An example endpoint in a docstring. | Yes; not contacted. |
| raw.githubusercontent.com | lib/cli/src/crewai_cli/constants.py:352; lib/crewai/src/crewai/constants.py:357; lib/devtools/src/crewai_devtools/litellm_catalog_update.py:20 | The LiteLLM model price catalog (JSON) for the CLI model picker and the maintainers' catalog refresh. | Expected: JSON data, not code (see the risk lines below). |
| ah-api.merge.dev | crewai_tools/tools/merge_agent_handler_tool/merge_agent_handler_tool.py:39 | Merge Agent Handler API. | Yes: a tool API. |
| api.brightdata.com | crewai_tools/tools/brightdata_tool/brightdata_dataset.py:13 | Bright Data dataset, SERP and unlocker APIs. | Yes: a tool API. |
| cognitiveservices.azure.com | crewai_tools/tools/azure_documentdb_vector_search_tool/vector_search.py:184 | An OAuth token scope for Azure OpenAI embeddings; an identifier, not a request target. | Yes. |
| a2ui.org | lib/crewai/src/crewai/a2a/extensions/a2ui/server_extension.py:23 | A2UI extension URIs and JSON schema ids; never fetched. | Yes. |
| api.serply.io | crewai_tools/tools/serply_api_tool/serply_job_search_tool.py:27 | Serply web, news, scholar, job and markdown APIs. | Yes: a tool API. |
| dashboard.oxylabs.io | crewai_tools/tools/oxylabs_base_tool/oxylabs_base_tool.py:123 | A sign-up link in docstrings. | Yes; not contacted. |
| a2a-protocol.org | lib/crewai/src/crewai/a2a/extensions/__init__.py:13 | A specification link in docstrings. | Yes; not contacted. |
| api.patronus.ai | crewai_tools/tools/patronus_eval_tool/patronus_eval_tool.py:13 | Patronus evaluation API. | Yes: a tool API. |
| developers.oxylabs.io | crewai_tools/tools/oxylabs_amazon_product_scraper_tool/oxylabs_amazon_product_scraper_tool.py:18 | Documentation links in docstrings. | Yes; not contacted. |
| docs.cloud.google.com | lib/crewai/src/crewai/rag/embeddings/providers/google/genai_vertex_embedding.py:90 | A migration link in deprecation warnings. | Yes; not contacted. |
| ai.dev | lib/cli/src/crewai_cli/constants.py:28 | Prompt text saying where to get a Gemini key. | Yes; not contacted. |
| api.cerebras.ai | lib/cli/src/crewai_cli/model_catalog.py:220; openai_compatible/completion.py:83 | Cerebras model listing and LLM base URL. | Yes: an LLM provider. |
| api.contextual.ai | crewai_tools/tools/contextualai_parse_tool/contextual_parse_tool.py:61 | Contextual AI parse, query and rerank APIs. | Yes: a tool API. |
| api.deepinfra.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:95 | DeepInfra LLM base URL. | Yes: an LLM provider. |
| api.deepseek.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:56 | DeepSeek LLM base URL. | Yes: an LLM provider. |
| opentelemetry.io | lib/crewai/src/crewai/telemetry/tracing/semantic_conventions.py:12 | Specification links in tracing modules. | Yes; not contacted. |
| actions.zapier.com | crewai_tools/adapters/zapier_adapter.py:10 | Zapier AI actions API. | Yes: a tool API. |
| api.stagehand.browserbase.com | crewai_tools/tools/stagehand_tool/stagehand_tool.py:166 | Default Stagehand API server. | Yes: a tool API. |
| dashscope-intl.aliyuncs.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:89 | Alibaba DashScope LLM base URL. | Yes: an LLM provider. |
| developers.google.com | crewai_tools/tools/serply_api_tool/serply_web_search_tool.py:49 | Reference links in field descriptions. | Yes; not contacted. |
| docs.apify.com | crewai_tools/tools/apify_actors_tool/apify_actors_tool.py:29 | Documentation links. | Yes; not contacted. |
| docs.hyperbrowser.ai | crewai_tools/tools/hyperbrowser_load_tool/hyperbrowser_load_tool.py:16 | Documentation links in a field description. | Yes; not contacted. |
| download.pytorch.org | lib/crewai/pyproject.toml:141,146 | uv package index for CPU torch wheels, used at install time. | Yes: a package source, not a run-time call. |
| infer.roboflow.com | lib/crewai/src/crewai/rag/embeddings/providers/roboflow/roboflow_provider.py:27 | Default Roboflow inference URL for the Roboflow embedder. | Yes: an embedding provider. |
| learn.microsoft.com | lib/crewai/src/crewai/llms/providers/azure/completion.py:1320 | A docstring link. | Yes; not contacted. |
| mcp.firecrawl.dev | scripts/wharf_runner.py:74 | The Firecrawl MCP server, used by a maintainer script that exercises MCP telemetry. | Yes: a maintainer script, not shipped code. |
| ossrdbms-aad.database.windows.net | crewai_tools/tools/azure_documentdb_vector_search_tool/auth.py:23 | An Entra ID token scope for DocumentDB authentication; an identifier. | Yes. |
| api.groq.com | lib/cli/src/crewai_cli/model_catalog.py:216 | The model picker lists Groq models. | Yes: an LLM provider. |
| api.parallel.ai | crewai_tools/tools/parallel_tools/parallel_search_tool.py:66 | Parallel search API. | Yes: a tool API. |
| api.weather.com | lib/crewai/src/crewai/project/crew_definition.py:175 | An example MCP reference in a docstring. | Yes; not contacted. |
| app.hyperbrowser.ai | crewai_tools/tools/hyperbrowser_load_tool/hyperbrowser_load_tool.py:25 | A sign-up link in a docstring. | Yes; not contacted. |
| bucket.r2.cloudflarestorage.com | crewai_tools/tools/url_read_tool/url_read_tool.py:178 | A presigned-URL example in a docstring. | Yes; not contacted. |
| cerebras.ai | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:150 | A provider link in a docstring. | Yes; not contacted. |
| cli.github.com | lib/devtools/src/crewai_devtools/cli.py:139 | An install hint printed by the release tool. | Yes; not contacted. |
| context7.com | context7.json:2 | Registers the repository with the Context7 documentation index; no code reads it. | Yes; not contacted. |
| dashscope.aliyun.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:151 | A provider link in a docstring. | Yes; not contacted. |
| deepinfra.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:152 | A provider link in a docstring. | Yes; not contacted. |
| deepseek.com | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:146 | A provider link in a docstring. | Yes; not contacted. |
| docs-platform.crewai.com | crewai_tools/adapters/enterprise_adapter.py:414 | A link in a legacy-token warning. | Yes; not contacted. |
| docs.astral.sh | lib/cli/src/crewai_cli/deploy/validate.py:866 | An install hint in deploy validation. | Yes; not contacted. |
| export.arxiv.org | crewai_tools/tools/arxiv_paper_tool/arxiv_paper_tool.py:28 | arXiv query API, over plain http. | Yes: a tool API (plain http is a weakness, not a sign of hostility). |
| google.serper.dev | crewai_tools/tools/serper_dev_tool/serper_dev_tool.py:117 | Serper search API. | Yes: a tool API. |
| httpbin.org | crewai_tools/tools/stagehand_tool/example.py:79 | An example target in an example script's prompt. | Yes; not contacted by the library. |
| mdb.ai | crewai_tools/tools/ai_mind_tool/ai_mind_tool.py:12 | AI Minds API base URL. | Yes: a tool API. |
| mindsdb.com | crewai_tools/tools/ai_mind_tool/ai_mind_tool.py:26 | A link in the tool description. | Yes; not contacted. |
| ollama.ai | lib/crewai/src/crewai/llms/providers/openai_compatible/completion.py:147 | A provider link in a docstring. | Yes; not contacted. |
| r.jina.ai | crewai_tools/tools/jina_scrape_website_tool/jina_scrape_website_tool.py:52 | Jina reader: fetches a page through Jina as text. | Yes: a tool API. |
| scrape.serper.dev | crewai_tools/tools/serper_scrape_website_tool/serper_scrape_website_tool.py:50 | Serper scrape API. | Yes: a tool API. |
| serpapi.com | crewai_tools/tools/serpapi_tool/serpapi_base_tool.py:47 | A link in the missing-key error message. | Yes; not contacted. |
| server.com | lib/crewai/src/crewai/agents/agent_builder/base_agent.py:383 | An example MCP URL in a field description. | Yes; not contacted. |
| stackoverflow.com | lib/crewai/src/crewai/project/crew_base.py:826 | A reference link in a comment. | Yes; not contacted. |
| unpkg.com | lib/crewai/src/crewai/flow/visualization/assets/interactive.js:42 | vis-network 9.1.2, loaded by the browser to draw the flow graph. | Expected; version-pinned but without subresource integrity. |
| www.bing.com | crewai_tools/tools/brightdata_tool/brightdata_serp.py:136 | A Bing search URL handed to Bright Data. | Yes; crewai does not connect to it. |
| www.stagehand.dev | crewai_tools/tools/stagehand_tool/example.py:76 | An example target in an example script's prompt. | Yes; not contacted by the library. |
| yandex.com | crewai_tools/tools/brightdata_tool/brightdata_serp.py:134 | A Yandex search URL handed to Bright Data. | Yes; crewai does not connect to it. |
Install hooks. The evidence found none: no package lifecycle script runs on install. The packages declare the stock hatchling build backend and ship as wheels on PyPI, so installing crewai runs no code from this repository. This is benign.
Risk-pattern lines. The evidence flags 43 lines: 6 download at run time from raw source hosting, 35 long encoded payloads, and 2 reads of credentials from the home directory. Three are outside tests, and each is the LiteLLM model price catalog (JSON data, not code) fetched from raw.githubusercontent.com/BerriAI/litellm/main.
crewai_cli/constants.py:352 is used at run time by provider.py:165 and model_catalog.py:405 to fill the CLI's model picker. The response is parsed as JSON and cached; it is never executed. Benign. The one weakness: it tracks a moving branch, so a bad upstream commit could show the picker wrong model names.crewai/constants.py:357 is exported, with no reader in the crewai package. Benign, and dead: the library ships its own checked-in copy of the catalog.crewai_devtools/litellm_catalog_update.py:20 is a maintainer tool that downloads the catalog with curl and writes it into the checkout as a reviewed PR before a release (update_catalog). Benign: maintainer-only, and the result is committed and reviewed.The other 40 lines are in tests and are all benign. The 35 long encoded payloads are base64 PDF, image and text attachments in recorded multimodal requests, plus Anthropic thinking-block signatures in recorded responses. The 2 home-directory credential reads are a fake C:/Users/alice/.ssh/id_rsa path that loader tests use to check a path is refused, not read. The 3 test-side raw-hosting downloads are image links inside a Firecrawl page recorded in a cassette. Every line has its verdict in the table below.
| Pattern | Where | Test | Verdict |
|---|---|---|---|
| Download at run time from raw source hosting | lib/cli/src/crewai_cli/constants.py:352 | no | Benign: the LiteLLM price catalog JSON for the CLI model picker, parsed and cached, never executed. |
| Download at run time from raw source hosting | lib/crewai/src/crewai/constants.py:357 | no | Benign: the same URL as an exported constant with no reader in the crewai package. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalAnthropic.test_pdf_file[anthropic-claude-3-5-haiku-20241022].yaml:4 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalAnthropic.test_pdf_file[anthropic-claude-3-5-haiku-20241022].yaml:109 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalFileTypes.test_pdf_anthropic.yaml:4 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalFileTypes.test_pdf_anthropic.yaml:109 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalFileTypes.test_text_gemini.yaml:4 | yes | Benign: a base64 text file ("Review Guidelines") in a recorded Gemini request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalFileTypes.test_text_gemini.yaml:75 | yes | Benign: a base64 text file in a recorded Gemini request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_mixed_files[gemini-gemini-2.0-flash].yaml:5 | yes | Benign: base64 image and text attachments in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_mixed_files[gemini-gemini-2.0-flash].yaml:77 | yes | Benign: base64 image and text attachments in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_mixed_files[gemini-gemini-2.5-flash].yaml:5 | yes | Benign: base64 image and text attachments in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_mixed_files[gemini-gemini-2.5-flash].yaml:73 | yes | Benign: base64 image and text attachments in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_text_file[gemini-gemini-2.0-flash].yaml:4 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_text_file[gemini-gemini-2.0-flash].yaml:74 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_text_file[gemini-gemini-2.5-flash].yaml:4 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestAgentMultimodalGemini.test_text_file[gemini-gemini-2.5-flash].yaml:70 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestCrewMultimodalAnthropic.test_pdf_file[anthropic-claude-3-5-haiku-20241022].yaml:9 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestCrewMultimodalFileTypes.test_pdf_anthropic.yaml:9 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestCrewMultimodalFileTypes.test_text_gemini.yaml:9 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestCrewMultimodalGemini.test_mixed_files[gemini-gemini-2.0-flash].yaml:10 | yes | Benign: base64 image and text attachments in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestCrewMultimodalGemini.test_text_file[gemini-gemini-2.0-flash].yaml:9 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestFlowMultimodalAnthropic.test_flow_with_pdf_file.yaml:9 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestFlowMultimodalGemini.test_flow_with_text_file.yaml:9 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/TestFlowMultimodalMultiStep.test_flow_with_mixed_files.yaml:10 | yes | Benign: base64 image and text attachments in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/TestAnthropicMultimodalIntegration.test_analyze_pdf.yaml:4 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/TestBedrockMultimodalIntegration.test_analyze_pdf.yaml:5 | yes | Benign: base64 PDF bytes in a recorded Bedrock request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/TestGeminiMultimodalIntegration.test_analyze_text_file.yaml:4 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/TestGenericFileIntegration.test_generic_file_pdf_anthropic.yaml:4 | yes | Benign: a base64 PDF attachment in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/TestGenericFileIntegration.test_generic_file_text_gemini.yaml:4 | yes | Benign: a base64 text file in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/anthropic/test_anthropic_thinking.yaml:43 | yes | Benign: a thinking-block signature in a recorded Anthropic response. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/anthropic/test_anthropic_thinking_blocks_preserved_across_turns.yaml:43 | yes | Benign: a thinking-block signature in a recorded response. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/anthropic/test_anthropic_thinking_blocks_preserved_across_turns.yaml:99 | yes | Benign: a thinking block replayed in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/llms/anthropic/test_anthropic_thinking_blocks_preserved_across_turns.yaml:139 | yes | Benign: a thinking-block signature in a recorded response. |
| Long encoded payload | lib/crewai/tests/cassettes/test_anthropic_thinking.yaml:43 | yes | Benign: a thinking-block signature in a recorded response. |
| Long encoded payload | lib/crewai/tests/cassettes/test_anthropic_thinking_blocks_preserved_across_turns.yaml:43 | yes | Benign: a thinking-block signature in a recorded response. |
| Long encoded payload | lib/crewai/tests/cassettes/test_anthropic_thinking_blocks_preserved_across_turns.yaml:99 | yes | Benign: a thinking block replayed in a recorded request. |
| Long encoded payload | lib/crewai/tests/cassettes/test_anthropic_thinking_blocks_preserved_across_turns.yaml:139 | yes | Benign: a thinking-block signature in a recorded response. |
| Reads credentials or keys from the home directory | lib/crewai/tests/project/test_crew_loader.py:767 | yes | Benign: a fake Windows path used to check the crew loader refuses a path outside the project; nothing is read. |
| Reads credentials or keys from the home directory | lib/crewai/tests/project/test_json_loader.py:85 | yes | Benign: the same fake path, checking the JSON loader refuses it. |
| Download at run time from raw source hosting | lib/crewai-tools/tests/tools/cassettes/firecrawl_search_tool_test/test_firecrawl_search_tool_integration.yaml:358 | yes | Benign: logo image links inside a recorded Firecrawl page. |
| Download at run time from raw source hosting | lib/crewai-tools/tests/tools/cassettes/firecrawl_search_tool_test/test_firecrawl_search_tool_integration.yaml:678 | yes | Benign: image links inside a recorded Firecrawl page. |
| Download at run time from raw source hosting | lib/crewai-tools/tests/tools/cassettes/firecrawl_search_tool_test/test_firecrawl_search_tool_integration.yaml:782 | yes | Benign: logo image links inside a recorded Firecrawl page. |
| Download at run time from raw source hosting | lib/devtools/src/crewai_devtools/litellm_catalog_update.py:20 | no | Benign: the maintainers' catalog refresh downloads the JSON with curl and commits it through a reviewed PR. |
| Function | Where | What it does |
|---|---|---|
_fetch_openai_compatible | lib/cli/src/crewai_cli/model_catalog.py:223 | Calls GET /models on an OpenAI-shaped API (OpenAI, Groq, Cerebras) with the user's key and keeps the chat models. Used by the CLI model picker. |
_fetch_anthropic | lib/cli/src/crewai_cli/model_catalog.py:241 | Lists Anthropic models from api.anthropic.com for the model picker, with display names and creation dates. |
_fetch_gemini | lib/cli/src/crewai_cli/model_catalog.py:257 | Pages through the Gemini models list, at most ten pages, keeping models that support content generation. |
_fetch_ollama | lib/cli/src/crewai_cli/model_catalog.py:313 | Lists the models installed on the local Ollama server, without a key. |
_fetch_litellm_data | lib/cli/src/crewai_cli/model_catalog.py:390 | Returns the cached LiteLLM catalog if fresh, otherwise downloads it once and caches it. On failure it falls back to a stale cache. |
fetch_provider_data | lib/cli/src/crewai_cli/provider.py:153 | Streams the LiteLLM catalog over HTTPS with certifi's CA bundle and writes it to the CLI's provider cache. |
get_latest_version_from_pypi | lib/crewai-core/src/crewai_core/version.py:112 | Asks pypi.org for crewai's latest non-yanked release, with a short timeout and a local cache, so the console can suggest an upgrade. |
_fetch_templates | lib/cli/src/crewai_cli/remote_template/main.py:154 | Pages through the crewAIInc organisation's public repositories on the GitHub API and keeps those named as templates. |
_download_zip | lib/cli/src/crewai_cli/remote_template/main.py:206 | Downloads the default-branch zipball of one template repository from the GitHub API. |
update_catalog | lib/devtools/src/crewai_devtools/litellm_catalog_update.py:28 | Downloads the LiteLLM catalog with curl, checks that it is a JSON object, and writes it into the checkout for a release PR. |
The exception strategy has three tiers, and they are applied consistently across the library. First, work failures are raised to the caller as typed Python exceptions: a guardrail that keeps failing raises ValueError, an agent that exceeds its time budget raises TimeoutError, an LLM that answers empty twice raises EmptyLLMResponseError. Second, before a failure is raised it is emitted as an event on the in-process event bus (TaskFailedEvent from the task, MethodExecutionFailedEvent from a flow method, AgentExecutionErrorEvent, CrewKickoffFailedEvent, LLMCallFailedEvent, ToolUsageErrorEvent), so console output, telemetry and tracing see it without catching it. Third, side channels never break a run: telemetry operations, trace export, background memory saves and event handlers catch their own errors, log or print them, and carry on. That last tier is deliberate and documented in the source ("telemetry must never break a command", "a trace is a record of the run, not a condition of it").
Retries are layered by what fails. The LLM layer retries only provider throttling; the executor retries an empty answer once; the agent re-runs a whole task on an unexpected error up to its retry limit; guardrails re-run a task until its output validates or their bound is reached; MCP retries tool discovery and tool calls. The bounds are the source defaults: max_iter 25, max_retry_limit 2, with guardrail_max_retries bounding guardrail re-runs and max_execution_time running the task under a timeout. Two exception types, ToolExecutionFailedError and HookAborted, are declared deliberate stops and pass straight through the retry loop.
| Mechanism | Where | Bound | What triggers it | When it is exhausted |
|---|---|---|---|---|
| LLM rate-limit retry | llms/retry.py:13-37, 124-205; llms/base_llm.py:246-288 | 3 attempts, 1 s initial delay, 8 s cap, 20% jitter, only for 429/throttling errors | An exception anywhere in the error chain with status 429, a throttling code (ratelimiterror, throttlingexception, resourceexhausted and similar) or a rate-limit phrase in its message. A provider Retry-After header replaces the computed delay and is not capped. A context variable stops a nested call from starting a second retry loop. Every BaseLLM subclass gets the wrapper on call and acall automatically. | The last error is re-raised unchanged; a non-throttling error is raised on the first attempt. ServiceQuotaExceededException is never retried. |
| Empty LLM answer retry | utilities/agent_utils.py:529-575 | 1 repeat | The model returns None, an empty string or no tool calls, or LiteLLM streaming raises EmptyLLMResponseError. The repeat is admitted through the RPM limiter like the first call. | EmptyLLMResponseError is raised to the agent. |
| Agent retry limit | agent/core.py:325-328, 908-960, 1074-1086 | max_retry_limit 2 (so up to 3 runs of a task) | Any exception from a task run that is not a timeout, a passthrough stop or a LiteLLM exception. Each failed attempt emits AgentExecutionErrorEvent and the whole execute_task runs again from the start. | The original exception is re-raised. LiteLLM exceptions are raised immediately without a retry. |
| Guardrail retries | task.py:290-292, 1425-1495; agent/core.py:396-398, 2260-2335 | guardrail_max_retries, default 3, on both Task and Agent | A guardrail (a callable or an LLM-judged string) rejects the output. The rejection text is appended to the conversation and the task or agent kickoff runs again. The deprecated Task max_retries field is copied into it with a warning. | ValueError naming the guardrail and its last error. |
| Execution timeout | agent/core.py:277, 1066-1128, 1243-1270 | max_execution_time seconds, off by default | The task run goes to a worker thread (sync) or an asyncio wait with a timeout (async) and the caller waits at most that long. | TimeoutError, emitted as AgentExecutionErrorEvent and raised without a retry. The sync worker thread is not interrupted; cancelling the future cannot stop a running thread, so the abandoned run continues in the background until it ends. |
| Iteration cap | agents/agent_builder/base_agent.py:285-287; utilities/agent_utils.py:374-445 | max_iter 25 | The reasoning loop has taken max_iter steps without a final answer. | One more LLM call with a "give your final answer" user turn; its output is accepted as the answer even when it does not parse. |
| RPM limiter | utilities/rpm_controller.py:12-90; utilities/agent_utils.py:463-472 | max_rpm per crew or agent, off by default | A model request when the per-minute counter has reached max_rpm. | The calling thread sleeps 60 s, then proceeds. It never fails a request. |
| SQLite busy timeout | flow/persistence/sqlite.py:26-27, 91-110; memory/storage/kickoff_task_outputs_storage.py:44-46 | 30 s per connection; schema creation retried every 0.05 s until the 30 s deadline | Another process or thread holds a lock on flow_states.db or latest_kickoff_task_outputs.db. Both run in WAL mode. The checkpoint SQLite provider (state/provider/sqlite_provider.py) also uses WAL but passes no timeout, so it waits the sqlite3 default of 5 s. | sqlite3.OperationalError is raised. |
| MCP operation retry | mcp/client.py:53-56, 153-260, 702-760 | 3 attempts with 1 s then 2 s waits; 30 s connect, discovery and execution timeouts | A tool listing or tool call that times out or fails with an error that is not authentication or not-found. There is no MCP connection retry: connect is bounded by its 30 s timeout and raises MCPConnectionError on failure, and the retry in _retry_operation wraps tool discovery and tool calls over an established connection (lines 425, 506, 638, 676). | ConnectionError naming the attempts. Authentication failures and not-found errors are raised on the first attempt. Tool calls are retried without any idempotency key. |
| Swallowed exceptions | telemetry/telemetry.py:235-252; events/event_bus.py:401-421; memory/unified_memory.py:355-368; events/listeners/tracing/trace_batch_manager.py:287-325 | 126 empty or comment-only error handlers in the evidence count | Telemetry spans, trace export, event handlers, background memory saves, TUI rendering and other best-effort paths. | Logged at debug or warning, printed to the console, or emitted as an event (MemorySaveFailedEvent); the run continues. |
Surfaces to the developer: every exhausted bound in the table above becomes an exception out of kickoff, after the matching failure event. A context window overflow is summarised away when respect_context_window is true (the Agent default); when it is false, handle_context_length raises SystemExit, which ends the host process rather than raising a catchable error. Tool errors are handled differently from other errors: a tool that raises is re-run up to its parsing-attempt limit (3, or 2 for some function-calling models), then the error text is given to the model as the tool's observation ("Moving on then") so the agent can choose another action. Only ToolExecutionFailedError stops the task.
Swallowed on purpose: telemetry (_safe_telemetry_operation returns None and logs at debug), trace export ("Failed to send trace events. Events will be lost." at warning; a refused AMP trace grant leaves the run untraced with a warning), event handler exceptions (_call_handlers prints them and continues, so a faulty listener never stops a crew), background memory saves (reported as MemorySaveFailedEvent, and drain_writes waits without re-raising). The evidence counts 126 empty or comment-only error handlers; the first it lists sit in the CLI's Textual TUIs and the checkpoint CLI, most marked noqa: S110 with the reason in a comment.
Absent: no dead-letter path (a failed run leaves its events, checkpoint and kickoff outputs locally and nothing queues it for retry), no circuit breaker (each call pays the full retry and timeout budget even when a provider is down), and no idempotency keys on LLM, tool or MCP calls, so the agent retry limit and the MCP retry can repeat a tool's side effect.
| Function | Where | What it does |
|---|---|---|
run_with_rate_limit_retry | llms/retry.py:124 | Runs an LLM call and repeats it up to three attempts when the error is a provider throttle, sleeping a jittered backoff between attempts. Any other error is raised at once. |
arun_with_rate_limit_retry | llms/retry.py:166 | The asyncio version of the rate-limit retry, awaiting the call and sleeping without blocking the event loop. |
get_retry_delay_seconds | llms/retry.py:100 | Returns the wait before a retry: the provider's Retry-After value when given, otherwise 1 s doubled per retry, capped at 8 s, with plus or minus 20% jitter. |
_retry_after_seconds | llms/retry.py:208 | Looks through an exception and its causes for a numeric Retry-After header in the shapes the provider SDKs use. |
_wrap_call_method | llms/base_llm.py:253 | Called from __init_subclass__ so that every LLM provider class has its own call and acall replaced by the rate-limit retry wrapper. |
call | llms/base_llm.py (abstract), each provider's completion.py | Sends a list of messages to the provider and returns its text or tool calls; each provider implements it, and the base class wraps it in the rate-limit retry. |
acall | llms/base_llm.py, each provider's completion.py | The asyncio counterpart of call, wrapped in the async rate-limit retry the same way. |
__init_subclass__ | llms/base_llm.py:246 | Runs when a BaseLLM subclass is defined and installs the retry wrapper on its call methods. |
_call_llm_retrying_empty | utilities/agent_utils.py:529 | Makes an LLM call and repeats it once if the answer is empty, admitting the repeat through the RPM limiter. Raises EmptyLLMResponseError if both are empty. |
_check_execution_error | agent/core.py:908 | Decides whether a failed task run may be retried: passthrough stops and LiteLLM errors are raised, otherwise it counts the attempt and raises once max_retry_limit is passed. |
_handle_execution_error | agent/core.py:936 | The sync retry path: checks the error, then calls execute_task again from the start. |
execute_task | agent/core.py:1003 | Runs one task for an agent, with or without a timeout, and sends any non-timeout failure to the retry handler. |
_execute_with_timeout | agent/core.py:1090 | Runs the task in a one-thread pool and waits up to max_execution_time seconds, converting a timeout to TimeoutError and other failures to RuntimeError. |
_aexecute_with_timeout | agent/core.py:1243 | The async version: awaits the task under an asyncio timeout and raises TimeoutError when the budget runs out. |
_process_kickoff_guardrail | agent/core.py:2260 | Validates a direct agent kickoff's output with its guardrail and, on rejection, appends the reason and re-runs, recursing until it passes or guardrail_max_retries is reached. |
_invoke_guardrail_function | task.py:1425 | Applies a task's guardrails to its output and re-executes the task with the rejection as feedback until it passes or the retry bound is hit. |
has_reached_max_iterations | utilities/agent_utils.py:374 | Returns true when the loop's iteration count has reached max_iter. |
handle_max_iterations_exceeded | utilities/agent_utils.py:387 | Asks the model one last time for a final answer and returns whatever it gives as the agent's finish. |
enforce_rpm_limit | utilities/agent_utils.py:463 | Calls the RPM limiter's admit function before a model request, if one is configured. |
check_or_wait | utilities/rpm_controller.py:38 | Admits a request while the minute's count is below max_rpm; otherwise sleeps 60 s and resets the count. Always returns true. |
init_db | flow/persistence/sqlite.py:91 | Creates the flow state tables, retrying every 0.05 s while the database is locked, up to the 30 s busy deadline. |
connect | mcp/client.py:153 | Opens the MCP transport and session within the connect timeout, raising MCPConnectionError on timeout or failure. |
_retry_operation | mcp/client.py:702 | Runs an MCP operation under a timeout and retries it with 1 s then 2 s waits, except for authentication and not-found errors, which are raised immediately. |
handle_context_length | utilities/agent_utils.py:1011 | On a context overflow, summarises the conversation when the agent respects the context window, otherwise raises SystemExit. |
kickoff | crew.py:1006 | Starts a crew run; any exception that escapes the task, agent and LLM layers is emitted as CrewKickoffFailedEvent and raised from here. |
_execute_core | task.py:850 | Runs a task through its agent, applies guardrails and output conversion, and emits TaskFailedEvent before re-raising any failure. |
_execute_method | flow/runtime/__init__.py:2927 | Runs one flow method (sync methods on a worker thread), emitting MethodExecutionFailedEvent on failure or a paused event when it raises HumanFeedbackPending. |
_safe_telemetry_operation | telemetry/telemetry.py:235 | Runs a telemetry span operation only when telemetry is enabled, and logs any exception at debug instead of raising it. |
_call_handlers | events/event_bus.py:401 | Calls every sync handler for an event, collects the handlers that raised, and prints their errors to the console without stopping the emitter. |
drain_writes | memory/unified_memory.py:355 | Blocks until every pending background memory save has finished, without re-raising their failures. |
Good. Retries are layered by failure type and each is bounded: LLM rate-limit retry with exponential backoff and jitter (3 attempts, 1 to 8 s), an empty-answer retry, agent max_retry_limit 2 and max_iter 25, guardrail retry bounds, max_execution_time timeouts, an RPM limiter, SQLite busy timeouts and schema retry, and MCP operation retry. Failures are emitted as events, and telemetry errors are dropped on purpose so they never break a run. The gaps are 126 empty handlers, no idempotency keys on LLM or tool calls, and no circuit breaker or dead-letter path; the technical assessment gives the reasoning per criterion.
There is no HA topology. CrewAI is a library and a CLI that run inside the developer's own process, so a crew or flow is exactly as available as the process, machine and network that host it; the repository deploys no servers, replicas, load balancers or failover (section 10). What the framework does provide is recovery inside a run's lifetime: checkpoints that let a crashed run resume, task replay from stored outputs, and flows that persist a paused state so a different process can pick them up later. The table names the three single points of failure first, then the three recovery mechanisms.
| Item | Kind | What depends on it, or what it recovers | Behaviour in the code |
|---|---|---|---|
| LLM provider | Single point of failure | Every agent step: each agent calls one configured model at one provider, and nothing in a run can proceed without it. | No failover to a second provider or model. LLM rate-limit retry: 3 attempts, 1 s initial delay, 8 s cap, 20% jitter, only for 429/throttling errors. Any other provider error, or an outage longer than the retries, fails the task and with it the crew or flow run. |
| Local storage directory | Single point of failure | Flow state, pending feedback, kickoff task outputs, memory and knowledge all live in one per-project directory under appdirs.user_data_dir on one disk (section 11). | No replica and no backup. Losing or corrupting the directory loses resumable flows, replay data and accumulated memory. The directory is keyed by the current directory's name unless CREWAI_STORAGE_DIR is set, so running from a renamed or different directory opens a new, empty set of stores. |
| CrewAI AMP platform | Single point of failure for platform features | CLI login, deploy, organisation and tool repository commands, and trace upload, all go to https://app.crewai.com by default. | A crew running locally does not need the platform: trace and telemetry failures are logged and dropped so they never stop a run. The CLI commands that talk to the platform fail until it returns. The availability of deployed crews is AMP's, outside this repository. |
| Checkpoint resume | Recovery mechanism | A crew, flow or agent that crashed or was stopped part-way. | Off unless a CheckpointConfig (or checkpoint=True) is set on the object or inherited from the crew or flow around it. It then writes the runtime state on each configured event, by default task_completed, to JSON files or SQLite. from_checkpoint restores the object so the next kickoff continues after the last completed task; fork restores onto a new branch. |
| Task replay | Recovery mechanism | Re-running a crew from a chosen task without repeating the tasks before it. | replay (also crewai replay -t) loads the latest kickoff's stored task outputs, feeds the earlier outputs back as context and runs from the chosen task. Only the most recent kickoff is kept, because each kickoff clears the table. |
| Human feedback resume | Recovery mechanism | A flow waiting for a person, across process restarts. | The flow raises HumanFeedbackPending after save_pending_feedback stores its state and pause context. Any later process rebuilds it with from_pending and continues with resume or resume_async, which clears the pending record. |
| Function | Where | What it does |
|---|---|---|
from_checkpoint | lib/crewai/src/crewai/crew.py:434 | Loads a checkpoint into a runtime state, finds the crew in it and restores its runtime, ready to resume from the last completed task on the next kickoff. The flow version is at lib/crewai/src/crewai/flow/runtime/__init__.py:656. |
fork | lib/crewai/src/crewai/crew.py:459 | Restores a crew from a checkpoint and moves the runtime state onto a new branch, so further checkpoints do not extend the original line. The flow version is at lib/crewai/src/crewai/flow/runtime/__init__.py:706. |
replay | lib/crewai/src/crewai/crew.py:2118 | Loads the stored outputs of the latest kickoff, validates the task to start from, and re-runs the crew from that task with the stored or supplied inputs. |
save_pending_feedback | lib/crewai/src/crewai/flow/persistence/sqlite.py:285 | In one transaction, appends a state snapshot and replaces the flow's pending feedback record with the pause context and state. |
from_pending | lib/crewai/src/crewai/flow/runtime/__init__.py:1260 | Builds a flow instance from the pending feedback record saved for a flow id, using the given persistence or the default one. |
resume | lib/crewai/src/crewai/flow/runtime/__init__.py:1349 | Continues a paused flow with the human's feedback, collapsing it to an outcome when the method declares outcomes, and triggers the downstream listeners. |
resume_async | lib/crewai/src/crewai/flow/runtime/__init__.py:1402 | The async version of resume, for callers already inside an event loop. |
Nothing in the repository backs up a store, and no RTO or RPO is stated anywhere: a search of the code, workflows and README for backup, RTO, RPO and disaster recovery finds only action names in the platform app catalogue. Restore exists only in the in-run sense above. The nearest thing to a recovery point is the checkpoint trigger, which is the last completed task by default and every event with on_events=["*"]; without checkpoints a crashed crew restarts from the beginning. Retention, and therefore how far back a restore can go, is unbounded unless max_checkpoints is set.
The release path has one safety step: a nightly .devYYYYMMDD pre-release of every package is published each day main changes, so a change can be installed and tried before a manual publish.yml release. There is no rollback in the repository: a bad release is replaced by publishing a new version, and users protect themselves by pinning. Rolling back a crew deployed to CrewAI AMP is a platform feature not visible here. There is no second region or account because there is no infrastructure; a developer's own deployment of a crew carries whatever disaster recovery its host provides. Nothing detects or alerts on a failed run (section 16): the developer sees the exception, or the failed trace in AMP when tracing is on.
Fair. As a library it has no HA topology of its own, but it does build recovery in: checkpoints with resume and fork, task replay from stored kickoff outputs, and flows that persist a pending human-feedback state and resume in a later process. Nothing backs up the local stores, no RTO or RPO is stated, every run depends on one LLM provider with no failover, and nothing detects or alerts on a failed run. The technical assessment gives the reasoning per criterion.
CrewAI runs inside the developer's process, so its performance is that process's performance plus the latency of every model and tool call it makes. Almost all wall-clock time in a run is spent waiting on the network: an agent completing one task makes between one and max_iter (25 by default) LLM calls, each sending the whole conversation so far, plus a tool call for most steps. The CPU work CrewAI does itself (prompt assembly, parsing, event emission, local SQLite and LanceDB writes) is small beside that. The scale questions that matter are therefore how many model calls a run makes, how many of them run at once, and how big the context sent on each one grows.
| Hot path | Entry function | Where | What it does | What drives its cost |
|---|---|---|---|---|
| Agent reasoning loop | invoke | agents/agent_executor.py:3132 | Runs the agent's reason, act, observe loop for one task until a final answer, returning a coroutine when called inside an event loop. AgentExecutor is itself a Flow, so each step is a flow method. | Steps up to max_iter; every step re-sends the full message history, so tokens per call grow with each tool observation until context-window summarisation trims them. The file is 3655 lines. |
| LLM call | get_llm_response | utilities/agent_utils.py:713 | Admits the request through the RPM limiter, runs the before-call hooks, calls the model (with the rate-limit retry wrapped round it) and repeats once on an empty answer. | Provider latency and token count. A throttled call can add up to 1 + 2 s of backoff, or the provider's Retry-After, before failing. |
| Memory recall and save | recall, remember_many | memory/unified_memory.py:686, 528 | recall waits for pending saves, then either embeds the query and runs one vector search (shallow) or runs a recall flow in which an LLM splits the query into sub-queries and searches scopes in parallel (deep, the default). remember_many queues records for the single background save worker. | Deep recall adds LLM calls on top of embedding calls; every save is an embedding call plus a LanceDB write, serialised one at a time, and recall blocks until the queue drains. |
| Knowledge retrieval | query | knowledge/knowledge.py:135 | Searches the agent's or crew's knowledge collections for the query and returns the top results above a similarity threshold (5 results, 0.6 by default). | One embedding call per query plus a ChromaDB vector search; ingestion chunks and embeds every source when the knowledge is first added. |
| Flow method dispatch | _execute_method | flow/runtime/__init__.py:2927 | Runs one flow method: awaits it if it is async, otherwise runs it on a worker thread, then persists state when the method is marked for persistence. | Thread hand-off per sync method and a SQLite write per persisted method; the runtime file is 4074 lines and kickoff_async alone is 431 lines. |
| Async task execution | execute_async | task.py:650 | Starts the task on a new daemon thread with a copy of the caller's context and returns a future the crew collects at the next synchronous task. | One OS thread per async task with no pool or upper bound; consecutive async tasks all run at once and all call the provider at once. |
Sync and asyncio APIs sit side by side on Crew, Flow and Agent. The sync API is the reference path; the async API either wraps it in a thread or runs natively. Inside a crew, tasks marked async_execution run on threads; inside a flow, listeners triggered by the same event run concurrently on the event loop. Memory saves go to a single background worker so writes never race. There is no process pool and no distributed execution: one run is one process.
| Function | Where | What it does | Concurrency it gives |
|---|---|---|---|
kickoff_async | crew.py:1146 | Runs the synchronous crew kickoff on a worker thread so an async caller can await it. | One thread per call; the crew inside is still sequential or hierarchical as configured. |
akickoff | crew.py:1226 | Native async kickoff that awaits tasks, memory and knowledge operations directly instead of wrapping the sync path in a thread. | Many crews can share one event loop without a thread each. |
_process_async_tasks | crew.py:2035 | Waits for the futures of the async tasks started since the last synchronous task and collects their outputs in order. | Joins the threads started by execute_async. |
_execute_listeners | flow/runtime/__init__.py:3171 | Finds the listeners and routers triggered by a finished method and runs them, gathering independent listeners together and racing the members of an OR group. | Concurrent listeners on one event loop; their sync bodies run on worker threads. |
_submit_save | memory/unified_memory.py:302 | Submits a memory save to the one-worker save pool and tracks its future, falling back to a synchronous save if the pool is shut down. | Saves run off the caller's thread but strictly one at a time. |
_execute_with_timeout | agent/core.py:1090 | Runs an agent's task in a one-thread pool and waits up to max_execution_time. | A thread per timed task; a timed-out run keeps its thread until it ends. |
Four caches exist, all in process and all scoped to one run or one process unless the developer configures otherwise. The tool result cache is opt-in: it is active only when the crew sets cache=True (the Crew default is false) or an agent is built with a cache handler. When active, the same tool with the same input string returns the stored output instead of calling again, and a declared tool failure is never stored; the field's own description warns against enabling it for live-data or state-changing tools. LanceDB, the memory store, is compacted in a background thread on start and every 100 saves so fragment files do not slow queries, and its scope column carries a BTREE index. The file upload cache stores provider file IDs for uploaded files so a file is not uploaded twice, in memory by default (Redis optional), with a 24 hour TTL (86,400 s) and 1000 entries. MCP tool lists can be cached per server for 300 s when a client is created with tool-list caching on (off by default). There is no cache of LLM responses.
| Cache | Where | What it does | Bound |
|---|---|---|---|
add (tool result cache) | agents/cache/cache_handler.py:23 | Stores a tool's output under the tool name and input string, skipping declared failures. | Unbounded dict for the life of the crew or agent; opt-in. |
read (tool result cache) | agents/cache/cache_handler.py:46 | Returns the stored output for a tool name and input, or nothing, under a read lock. | As above. |
_compact_async | memory/storage/lancedb_storage.py:238 | Starts a daemon thread that runs LanceDB's table optimise to merge fragment files, logging any failure at debug. | Every 100 saves and once at start. |
| Limit | Value | Where | When it bites |
|---|---|---|---|
| Iteration cap | max_iter 25 | agents/agent_builder/base_agent.py:285-287 | A task that needs more tool steps than this is forced to a final answer by one extra LLM call once the count is reached. It is also the only bound on a single task's cost: 25 calls with a long history is the worst case per attempt, times up to 3 attempts under max_retry_limit 2. |
| Requests per minute | max_rpm, unset by default | utilities/rpm_controller.py:38-70 | When set, the request that crosses it sleeps a full 60 s regardless of how far into the minute it is. Unset, nothing stops a crew with parallel async tasks from hitting the provider's own quota. |
| Provider throttling | 3 attempts, 1 s initial delay, 8 s cap, 20% jitter (with 3 attempts the waits are about 1 s and 2 s, so the cap is never reached) | llms/retry.py:35-38 | Sustained 429s from the provider: after the third throttled attempt the error reaches the agent, which may re-run the whole task. |
| Context window | the model's window; summarisation on by default | utilities/agent_utils.py:1011-1046; agent/core.py:317-320 | Long tool outputs or many steps overflow the window; each overflow costs an extra summarisation pass of LLM calls, and with summarisation off the process exits. |
| Memory scan limit | 50,000 rows | memory/storage/lancedb_storage.py:34, 486-512 | Listing, category and scope operations that scan the memories table without a vector query read at most 50,000 rows after filtering. |
Load evidence: none. There is no benchmark, load test, profiling harness or stated load level anywhere in the repository; CI runs pytest with a 60 s per-test timeout, which bounds test duration but measures nothing. How many concurrent crews one process sustains, or how memory recall degrades with store size, is not recorded.
Fair. The concurrency model is deliberate (sync and asyncio entry points, threaded async tasks, a background memory-save worker, concurrent flow listeners), with a tool result cache, a LanceDB scope index and compaction, an RPM limiter and context-window summarisation. The gaps are that there is no benchmark, load test or stated load level anywhere, the agent loop is bounded only by max_iter and provider limits, and the hot paths sit in files of over 3,000 lines. The technical assessment gives the reasoning per criterion.
CrewAI is a library, so its observability is designed for two audiences: the developer watching a run, and the CrewAI platform collecting what the developer agrees to send. Everything starts on the in-process event bus. Every crew, task, agent, LLM call, tool use, memory operation and flow method emits typed events; one set of listeners turns them into console output, another into anonymous telemetry spans, another into execution traces for the CrewAI AMP trace viewer. Nothing in the repository ships logs, metrics or alerts to an operations system, because the library deploys nothing of its own; a developer who hosts a crew wires those into their own application.
The table reproduces the evidence's six signals in full, with its present flag and match counts, then says what each one actually is after reading the matches.
| Signal | Evidence says present | Matches | Files | What it is | Where it goes and what it covers |
|---|---|---|---|---|---|
| Logging | yes | 593 | 122 | Present. Standard library module-level loggers (one per module, named after it), mostly at debug and warning, plus a coloured console printer gated by each agent's or crew's verbose flag. | To whatever handlers the host application configures; the library adds none except the A2A module's opt-in JSON formatter. Debug lines cover swallowed telemetry, tracing and storage failures; the console printer covers agent steps, tool calls and retries as they happen. |
| Metrics | yes | 2 | 1 | False positive. Both matches are an event emission sequence counter in events/base_events.py:18-39 that orders events within a context, not a metric. | Nowhere. Token usage is collected per run and returned on the crew output as token_usage, but no metrics backend, counter export or dashboard exists. |
| Tracing | yes | 286 | 57 | Present, in two forms. Anonymous telemetry is on by default and exports OpenTelemetry spans to https://telemetry.crewai.com:4319; it is off when OTEL_SDK_DISABLED, CREWAI_DISABLE_TELEMETRY or CREWAI_DISABLE_TRACKING is set. Execution traces are sent in batches to the CrewAI platform through PlusAPI (initialize_trace_batch, send_trace_events and ephemeral variants), controlled by CREWAI_TRACING_ENABLED and crewai traces enable/disable. | Telemetry goes to CrewAI's collector and records that crews, agents, tasks and tools ran, without prompts or outputs unless the crew sets share_crew (false by default). Execution traces go to the developer's AMP account (or an anonymous ephemeral batch with an access code) and cover the full event stream of a run, viewable as a timeline at the link printed when the run ends. A newer path in execution.py asks AMP for a trace grant and exports OpenTelemetry spans to the collector URL the grant names. |
| Error reporting | yes | 97 | 22 | False positive. The matches are Sentry hosts inside Content-Security-Policy headers recorded in HTTP test cassettes under lib/crewai/tests/cassettes/. | Nowhere. No error reporting SDK is installed or called; failures surface as exceptions to the developer and as failure events on the bus. |
| Alerting | yes | 5 | 2 | False positive. The matches are PagerDuty as a platform app name and its action names in crewai_core/platform_app_defaults.py:274 and platform_apps.py:314, a tool an agent can call, not an alert route. | Nowhere. Nothing in the repository pages or notifies anyone. |
| Health checks | no | 0 | 0 | Absent. The library runs no long-lived service of its own (the A2A server and the deployed crew on AMP are hosted by other code), so there is no endpoint to probe. | Nowhere. |
| Function | Where | What it does |
|---|---|---|
setup_listeners | events/event_listener.py:177 | Registers the default console and telemetry handlers for every event type on the bus. At 904 lines it is the longest function in the repository. |
register_handlers | telemetry/tracing/handlers.py:132 | Subscribes the span-building handlers that turn crew, agent, task, LLM and tool events into OpenTelemetry spans for a trace session. |
_safe_telemetry_operation | telemetry/telemetry.py:235 | Runs a telemetry span operation only when telemetry is enabled and drops any error with a debug log. |
_start_tracing | execution.py:154 | At the start of a run, decides whether to trace it and either opens an authenticated AMP trace grant, an anonymous ephemeral trace, or nothing; a refused grant leaves the run untraced with a warning. |
initialize_trace_batch | crewai_core/plus_api.py:599 | Asks the CrewAI platform to open a trace batch for a run and returns its response. |
send_trace_events | crewai_core/plus_api.py:616 | Posts a batch of collected trace events to the platform for an open trace batch. |
_send_events_to_backend | events/listeners/tracing/trace_batch_manager.py:287 | Sends the buffered events for the current batch through PlusAPI, logging "Events will be lost" at warning when the send fails. |
_finalize_backend_batch | events/listeners/tracing/trace_batch_manager.py:386 | Closes the batch on the platform and prints a console panel with the link to the trace viewer. |
_get_or_create_counter | events/base_events.py:18 | Returns the per-context counter that numbers emitted events, the match the evidence read as a metric. |
get_next_emission_sequence | events/base_events.py:33 | Takes the next number from that counter and records it as the last emitted sequence. |
The developer, at the console, immediately, when they ran the crew themselves: an exception reaches them from the kickoff call and verbose mode shows each step, retry and tool error as it happens. The developer, later, in the AMP trace viewer, if tracing is enabled and they open the printed link. For a crew deployed to CrewAI AMP, detection belongs to the platform, which is outside this repository. Nobody, otherwise. A crew embedded in someone's service fails into that service's own logs and error handling; nothing in this repository pages anyone, counts failures or exposes a health endpoint. Swallowed side-channel failures (telemetry, trace export, background memory saves) are visible only at debug or warning level, so a trace batch that never arrives is noticed when someone looks for it. Logs exist and the library reads none of them back; that is a design fact for a library, and the host application owns the rest.
The evidence lists 193 duplicate-body groups and 84 unreferenced functions as candidates. Each was checked against the source; 22 items survive. Eight are copies of a body that already exists in a class the copy could inherit from or a helper it could share. Fourteen are private functions whose name appears exactly once in the repository, at the definition: a whole-word search of every Python file under lib/, scripts/ and the root conftest.py finds no call, no string lookup and no re-export. Because they are private, no other repository should be calling them either. Deleting any of the 22 changes nothing the system does.
| Code | File and lines | What it does | What makes it redundant |
|---|---|---|---|
_chunk_text (six overrides) | crewai/knowledge/source/csv_knowledge_source.py:46-51, excel_knowledge_source.py:176-181, json_knowledge_source.py:58-63, pdf_knowledge_source.py:58-63, string_knowledge_source.py:36-41, text_file_knowledge_source.py:38-43 | Splits a source's text into chunk_size slices that overlap by chunk_overlap characters, ready for embedding. | A copy of BaseKnowledgeSource._chunk_text (base_knowledge_source.py:43-48), which every one of the six classes already inherits; each override is identical to it. |
ExcelKnowledgeSource: validate_file_path, _process_file_paths, validate_content | crewai/knowledge/source/excel_knowledge_source.py:31-43, 45-72, 74-89 | Validate that a file path was given, normalise the deprecated file_path and file_paths fields into a list of paths, and log any path that does not exist. | Copies of the same three methods on BaseFileKnowledgeSource (base_file_knowledge_source.py:30-42, 89-116, 54-69). The Excel source subclasses BaseKnowledgeSource instead of BaseFileKnowledgeSource, as the CSV, JSON, PDF and text sources do, and so re-implements them. |
create_crew_from_zip (crewai_cli) | cli/src/crewai_cli/plus_api.py:167-193 | Uploads a local project ZIP archive to the CrewAI platform to create a crew deployment, with an optional name, environment and project id. | The crewai_cli PlusAPI subclasses the crewai_core PlusAPI, and this override is identical to the parent method it replaces (crewai_core/plus_api.py:528-554). |
ContextWindowSummarizationEvent.__init__ | crewai/events/types/llm_events.py:131-144 | Replaces the task and agent objects passed to the event with their id, name and role before the event is built, so the event does not hold live objects. | Identical to LLMEventBase.__init__ in the same file (lines 15-28). The event cannot simply subclass LLMEventBase, which requires a call_id, but the shared body belongs in one place both inherit from. |
get_client_id (five copies) | crewai_core/auth/providers/auth0.py:30-35, entra_id.py:32-37, keycloak.py:26-31, okta.py:30-35, workos.py:26-31 | Returns the OAuth2 client id from the provider settings, or raises ValueError when it is not configured. | Five identical bodies in the five providers; the abstract method on BaseProvider could carry the body once. |
_initialize_firecrawl (two copies) | crewai_tools/tools/firecrawl_crawl_website_tool/firecrawl_crawl_website_tool.py:82-105, firecrawl_search_tool/firecrawl_search_tool.py:81-104 | Creates the Firecrawl SDK client with the tool's API key, offering to install firecrawl-py with uv if it is missing. | The two tools carry the same 24-line body; one shared helper (or a common Firecrawl base) would serve both. |
_get_file_path (two copies) | crewai_files/uploaders/bedrock.py:52-64, gemini.py:75-87 | Returns the filesystem path of a file input when its source is a local path, otherwise nothing. | The Bedrock and Gemini uploaders define the same module-level function; it belongs in the shared uploader base module. |
| lib/crewai/tests/cli (23 Python files) | lib/crewai/tests/cli/ against lib/cli/tests/ | Tests the CLI's OAuth2 providers, token manager, settings, version checks and PlusAPI client. | The largest duplicate groups in the evidence are this suite and lib/cli/tests line for line (Okta 251 lines, AtomicFileOperations 145, EntraId 135, Keycloak 132, Settings 131, TokenManager 124, and more). Both exercise the same crewai_core code; the copy under lib/crewai reaches it through the crewai re-export modules (crewai.auth, crewai.plus_api, crewai.settings, crewai.version, documented as stable import paths), and 8 of its files import crewai_cli directly. None of its 23 files imports the deprecated crewai.cli shim. |
_refused | cli/src/crewai_cli/experimental/eval_crew.py:2398 | Prints the refusal message for an AMP HTTP response and exits with status 1. | No caller anywhere in the repository. |
_has_lockfile | cli/src/crewai_cli/run_crew.py:531 | Returns true when the project has a uv or Poetry lockfile. | No caller; the run code calls the uv and Poetry lockfile checks it combines directly (run_crew.py:550, 565). |
_next_selectable_index | cli/src/crewai_cli/tui_picker.py:477 | Moves a picker cursor one step in a direction, skipping separator rows, and stays put at either end. | No caller anywhere in the repository. |
_handle_api_response | crewai_tools/tools/scrapegraph_scrape_tool/scrapegraph_scrape_tool.py:143 | Validates a Scrapegraph API response, raising a rate-limit or runtime error when it is empty, carries an error or lacks a result. | No caller; the tool never routes its responses through it. |
_crew_to_agent_card | crewai/a2a/utils/agent_card.py:446 | Builds an A2A AgentCard describing a crew's name, description and skills at a given URL. | No caller anywhere in the repository. |
_normalize_grpc_metadata | crewai/a2a/utils/delegation.py:656 | Lowercases every gRPC metadata key, because gRPC rejects mixed-case header names. | No caller; the gRPC interceptors beside it do not use it. |
_extract_tool_name | crewai/agents/agent_executor.py:2482 | Reads and sanitises the tool name from the several shapes a provider's tool call can take. | No caller anywhere in the repository. |
_supports_stop_words_implementation | crewai/llms/base_llm.py:670 | Returns true when the LLM instance has stop sequences configured. | Its docstring says native providers can return it from their stop-word check, but none does. |
_ahandle_tool_use_conversation | crewai/llms/providers/anthropic/completion.py:2004 | Runs the async Anthropic tool-use exchange: executes the requested tools and sends their results back for a final answer. | No caller anywhere in the repository. |
_is_nova_model | crewai/llms/providers/bedrock/completion.py:2284 | Returns true when the Bedrock model id names an Amazon Nova model. | No caller anywhere in the repository. |
_get_document_format | crewai/llms/providers/bedrock/completion.py:2323 | Maps a document MIME type to the Bedrock document format name (pdf, csv, docx and others). | No caller anywhere in the repository. |
_get_video_format | crewai/llms/providers/bedrock/completion.py:2345 | Maps a video MIME type to the Bedrock video format name (mp4, mov, webm and others). | No caller anywhere in the repository. |
_is_async_callable | crewai/tools/base_tool.py:113 | Returns true when a callable is a coroutine function. | No caller; a one-line wrapper over the standard library check it calls. |
_call_llm_with_reasoning_prompt | crewai/utilities/reasoning_handler.py:610 | Sends a reasoning-plan prompt for a task to the LLM and returns the plan text. | Marked deprecated and "kept for backward compatibility", but it is private and nothing calls it; superseded by the current reasoning handler. |
Most of the evidence's candidates are not redundant. Of the unreferenced functions, the public ones and those reached by name were dropped. That covers the Textual action_ and on_ handlers in the CLI TUIs, which Textual calls through key bindings and message names, and the gRPC intercept_ methods and init_poolmanager, which are framework overrides. It also covers create_module (the importlib loader protocol behind the deprecated crewai.cli shim), the conftest iter_ shims and pytest_recording_configure (an aiohttp stream interface and a pytest hook), and public API such as copy_template_files and the a2a helpers, which other repositories may call. Of the duplicate groups, the test-only ones other than the CLI suite were dropped: repeated fixtures and helper functions inside test modules are a test-style choice, not dead code. The small per-class protocol methods were dropped too, because lifting them would add a base class for four to seven lines. One family shows the shape of the dropped items:
| Candidate | Where | What it does | Why it is not redundant |
|---|---|---|---|
action_resume | cli/src/crewai_cli/checkpoint_tui.py:716 | Closes the checkpoint browser with "resume" as the chosen action for the selected checkpoint. | Textual calls action_ methods by name from key bindings. |
on_tree_node_selected | cli/src/crewai_cli/memory_tui.py:273 | Loads and lists the memory entries for the scope the user selected in the memory browser's tree. | Textual dispatches messages to on_ methods by name. |
intercept_unary_unary | crewai/a2a/utils/delegation.py:706 | Injects the A2A auth metadata into a unary gRPC call before it is sent. | A gRPC client interceptor override; gRPC calls it. |
init_poolmanager | crewai_tools/security/ssrf_adapter.py:271 | Builds the connection pool for the SSRF-validating HTTP adapter. | A requests adapter override; requests calls it when the adapter is mounted. |
create_module | crewai/cli/__init__.py:36 | Returns the already-imported crewai_cli module for an old crewai.cli import path. | Part of the importlib loader protocol; the import system calls it. |
iter_chunked | conftest.py:45 | Iterates a recorded response body in chunks of a given size. | Completes the aiohttp stream interface for replayed cassettes; aiohttp clients call it. |
pytest_recording_configure | conftest.py:448 | Registers the custom VCR request matchers (a Bedrock host matcher) for each cassette session. | A pytest plugin hook, called by name. |
copy_template_files | cli/src/crewai_cli/create_crew.py:162 | Copies the project template files into a new crew project. | Public API of crewai_cli that other code may import. |
is_retryable_error | crewai/a2a/errors.py:464 | Says whether an A2A error code (internal error, rate limit exceeded, task timeout) might succeed on retry. | A public helper in the a2a package that code building on the A2A integration may import. |
__aexit__ (four copies) | crewai/mcp/client.py:788, transports/http.py:187, sse.py:105, stdio.py:162 | Disconnects the MCP client or transport when an async with block ends. | The async context manager protocol, implemented by each class over its own disconnect. |
get_audience (three copies) | crewai_core/auth/providers/auth0.py:23, entra_id.py:25, okta.py:23 | Returns the configured OAuth2 audience, or raises when it is missing. | Only three of five providers share this body (WorkOS and Keycloak return a fallback instead of raising), so there is no single place for it, unlike get_client_id. |
_get_sync_client (four copies) | crewai/llms/providers/anthropic/completion.py:339, azure/completion.py:373, bedrock/completion.py:352, openai/completion.py:479 | Builds the provider's SDK client on first use and returns the cached one after that. | A four-line lazy accessor over each provider's own client builder. |
unwrap (three copies) | crewai/flow/flow_wrappers.py:104, project/wrappers.py:219, 327 | Returns the original method a decorator wrapper holds. | A one-line accessor on three unrelated wrapper classes. |
A diagram earns its place here only when it shows something the source cannot, usually a relationship that crosses a boundary the call graph cannot follow. On that test this document draws ten figures: the module dependency figure in section 4; the layered call graph in section 5, with dashed edges for calls the graph does not resolve; four sequence diagrams in section 6 (the crewai run subprocess, A2A delegation, human feedback pause and resume, telemetry and trace export); three state diagrams in section 7 (the AgentExecutor reasoning loop, the Flow human feedback lifecycle, the A2A task lifecycle); and the deployment topology in section 10. Everything else was considered and left out, for the reasons below.
| Diagram type rejected | Why it was not drawn | Where the same ground is covered |
|---|---|---|
| class diagram | The Pydantic model hierarchy (Agent, Task, Crew, Flow, BaseTool, BaseLLM) is readable from the source and the 552-file package would make an unreadable figure; the call graph and state machines show the relationships that matter. | Sections 5 and 7 |
| entity-relationship diagram | Every store is an independent single-purpose local file with one or two tables and no foreign keys between them; the store and index tables in section 11 say everything an ER diagram would. | Section 11 |
| control-flow flowchart | Ordinary control flow inside kickoff or the agent loop is what the code already is; the genuine branching is drawn as state machines in section 7. | Section 7 |
| network diagram | The repository deploys no infrastructure: no VPC, hosts, containers or IaC, so there is no network to draw beyond the deployment topology in section 10. | Section 10 |
| per-tool integration diagram | The 84 tools each make one outbound call to one API; a fan of 84 arrows says less than the integrations and outbound calls tables in section 12. | Section 12 |
Every absence the sections state, one line each, with the section that states it. This document describes; it does not judge. Which of these absences is a risk, how serious it is, what it would cost to operate around and what to fix first are questions for the technical assessment, which rates each one in its risk register (section 11) and orders the fixes in its remediation roadmap (section 12).
authenticate in crewai, and EnterpriseTokenAuth is not implemented (section 8)..env.example, no single configuration reference and no central configuration schema (section 9).task_index in the task outputs table (section 11).