feat: real CrewAI integration using mcps field (not simulation)

[ascender1729]

Summary

Adds integrations/examples/crewai_real_integration.py, a real CrewAI integration that replaces the simulation-only crewai_compliance.py pattern.

The new script uses the actual crewai.Agent / Task / Crew / Process classes, and connects to the Attestix MCP server over stdio via crewai_tools.MCPServerAdapter (which spawns python -m main and exposes all 47 Attestix tools as native CrewAI BaseTool wrappers).

Real vs simulation

Aspect	crewai_compliance.py (existing)	crewai_real_integration.py (new)
Imports	Local @dataclass AgentRole stubs	from crewai import Agent, Task, Crew, Process
Tool host	In-process service calls	MCPServerAdapter + stdio JSON-RPC subprocess
Tool wrappers	None	47 real CrewAIMCPTool objects on Agent(tools=...)
Agent type check	N/A	assert isinstance(officer, crewai.Agent)
MCP transport	None	mcp.StdioServerParameters(command='python', args=['-m','main'])
LLM handling	N/A	Stub LLM when no OPENAI_API_KEY, real LLM when set

Crew layout

1. Researcher - maps system to EU AI Act Annex III
2. Compliance Officer - holds all 47 Attestix MCP tools, drives compliance checkpoints
3. Report Writer - assembles final artefact

Compliance checkpoints exercised through the real CrewAI MCP tools

A. create_agent_identity for every crew member
B. create_compliance_profile (high-risk agent uses Annex III category 3 so Article 43 self-assessment is permitted)
C. log_action for every crew task (hash-chained audit trail)
D. record_conformity_assessment + generate_declaration_of_conformity for the high-risk Compliance Officer

Why not crew.kickoff()

Driving the Attestix MCP tools directly through the CrewAI wrappers keeps the example deterministic and free of external LLM dependencies, while still exercising the full CrewAI tool wrapper -> mcpadapt -> stdio JSON-RPC -> Attestix MCP server -> services -> on-disk storage code path. If OPENAI_API_KEY is set, CrewAI uses the real LLM; otherwise a stub LLM is installed purely to let Agent construction succeed (it is never invoked).

Simulation fallback

If crewai, crewai-tools, or mcp are not installable, or the real path raises at runtime, the script falls back to the same in-process flow used by google_adk_compliance.py. The fallback is clearly labelled in the output.

Test plan

• pip install crewai 'crewai-tools[mcp]' completes successfully
• python integrations/examples/crewai_real_integration.py exits 0
• Output shows Connected. 47 Attestix MCP tools exposed as CrewAI tools.
• Output shows isinstance(*, crewai.Agent): True for all three agents
• Output shows crew type: crewai.crew.Crew with Process.sequential
• 3 Attestix identities persisted to ~/.attestix/identities.json
• 6 hash-chained audit entries persisted to ~/.attestix/provenance.json
• Declaration of conformity issued (decl:...) for the high-risk agent
• All 15 existing tests/integration/test_framework_integrations.py tests still pass

Reuse snippet

from crewai import Agent
from crewai_tools import MCPServerAdapter
from mcp import StdioServerParameters

params = StdioServerParameters(command='python', args=['-m', 'main'])
with MCPServerAdapter(params) as attestix_tools:
    officer = Agent(role='Compliance Officer', tools=attestix_tools, ...)

[coderabbitai]

Warning

Rate limit exceeded

@ascender1729 has exceeded the limit for the number of commits that can be reviewed per hour. Please wait 8 minutes and 42 seconds before requesting another review.

Your organization is not enrolled in usage-based pricing. Contact your admin to enable usage-based pricing to continue reviews beyond the rate limit, or try again in 8 minutes and 42 seconds.

⌛ How to resolve this issue?

After the wait time has elapsed, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

We recommend that you space out your commits to avoid hitting the rate limit.

🚦 How do rate limits work?

CodeRabbit enforces hourly rate limits for each developer per organization.

Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout.

Please see our FAQ for further information.

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: a9ef9e08-af10-4fac-914b-8e218f1df8fa

📥 Commits

Reviewing files that changed from the base of the PR and between d3e9a69 and 59725f3.

📒 Files selected for processing (1)

• integrations/examples/crewai_real_integration.py

Note

.coderabbit.yaml has unrecognized properties

CodeRabbit is using all valid settings from your configuration. Unrecognized properties (listed below) have been ignored and may indicate typos or deprecated fields that can be removed.

⚠️ Parsing warnings (1)

Validation error: Unrecognized key(s) in object: 'ignore'

⚙️ Configuration instructions

• Please see the configuration documentation for more information.
• You can also validate your configuration using the online YAML validator.
• If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/real-crewai-integration

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}