[BUG] Fix #511: BFT Auditor consensus hangs indefinitely on provider timeout

[adityakryadav]

Summary

Closes #511.
When one or more auditor providers stalled indefinitely, the asyncio.gather() call in the consensus logic would hang forever, blocking the entire audit pipeline.

Changes

• agentwatch/auditors/bft_auditor.py (new file) - Introduces BFTAuditor:
  • audit_step() wraps asyncio.gather() with asyncio.wait_for(..., timeout=1.5).
  • On asyncio.TimeoutError, returns {"error": "Global consensus timeout", "consensus_score": 0.0} immediately instead of hanging.
  • On quorum failure (too many providers returned exceptions), returns {"error": "Quorum not reached", ...}.
• agentwatch/auditors/__init__.py (new file) - Exports BFTAuditor.

How to test

import asyncio
from agentwatch.auditors import BFTAuditor
auditor = BFTAuditor(configs=[{"provider": "a"}, {"provider": "b"}])
result = asyncio.run(auditor.audit_step(1, {"tool": "bash"}))
# Should complete within 2s and return consensus_score
print(result)

[coderabbitai]

Warning

Review limit reached

@adityakryadav, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 58 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: c04ef421-91a2-46d5-b3e7-8c09274fffd4

📥 Commits

Reviewing files that changed from the base of the PR and between a871b5a and 37a6f9d.

📒 Files selected for processing (3)

• agentwatch/auditors/__init__.py
• agentwatch/auditors/bft_auditor.py
• tests/test_multiagent.py

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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.}

[github-actions]

🧪 PR Test Results

Check	Result
Tests (pytest tests/)	✅ success
Lint (ruff check .)	✅ success
Coverage (agentwatch)	73.51%

_{Python 3.12 · commit 37a6f9d}

[vercel]

@adityakryadav is attempting to deploy a commit to the sreerevanth's projects Team on Vercel.

A member of the Team first needs to authorize it.

[sreerevanth]

Thanks for the contribution! I like the idea of preventing indefinite hangs. Before we merge this, I'd like to keep the implementation aligned with the existing auditing architecture rather than introducing a new BFT auditor implementation. Could you update the existing consensus logic to add timeout handling, make the timeout configurable, and include a regression test that simulates a stalled provider? That would make the fix much easier to integrate.