CodeRabbit auto-review never fires: inherited label gate deadlocks against auto_apply_labels

[groupthinking]

Summary

CodeRabbit auto-review is skipped on every pull request in this repository. The CodeRabbit commit status reads:

Review skipped: excluded by label configuration

The cause is a required-labels gate inherited from the CodeRabbit dashboard / org UI. It is not present in .coderabbit.yaml, so it is invisible in the repo and has never been reviewed in a PR.

The deadlock

The inherited gate requires at least one of ~26 labels before a review will start:

architecture-gap, bug, ci-cd, ci/cd, copilot-rabbit, documentation, duplicate,
enhancement, frontend, github_actions, good first issue, help wanted,
high-priority, invalid, javascript, ml-model, needs-triage, pipeline-critical,
placeholder-code, priority:high, python, python:uv, question, styling, tests, v0

Four of these — architecture-gap, ci-cd, pipeline-critical, placeholder-code — are labels CodeRabbit itself applies, as part of a review, via reviews.auto_apply_labels: true and reviews.labeling_instructions in .coderabbit.yaml.

A pull request opens unlabelled. The gate therefore skips it, and because the review never runs, nothing applies a label that would un-skip it. Labelling after the fact does not retro-trigger the review.

Evidence

Both currently-open non-draft PRs, on their current heads:

PR	Head	CodeRabbit status
#1381	8628b5e	Review skipped: excluded by label configuration
#1410	3f72b75	Review skipped: excluded by label configuration

#1410 carries documentation and ci-cd, both of which are in the required list — and it is still skipped, because the labels were added after open. This is recorded on the PR itself:

Auto-review skipped at open time for want of a qualifying label; adding documentation and ci-cd afterwards did not retro-trigger it, so requesting explicitly.

Every CodeRabbit review this repo has received was requested by hand with @coderabbitai full review.

Why it matters

Any automated PR-remediation loop that expects CodeRabbit to review on open silently gets nothing. The failure is quiet — the commit status is success, not a failure, so a green CodeRabbit status currently means "did not review" just as often as it means "reviewed and found nothing".

Secondary finding

.coderabbit.yaml's header claims to be the single source of truth for settings that previously lived only in the dashboard. This gate was not in it. With inheritance: true, any setting the file leaves unset can be overridden from the dashboard without review — so "unset" is not a safe way to express "I want the default".

Acceptance criteria

• reviews.auto_review.labels is set explicitly in .coderabbit.yaml so the inherited gate no longer applies
• Draft PRs remain excluded (drafts: false unchanged)
• No other CodeRabbit setting changes
• The header comment records that unset values are dashboard-overridable

[linear-code]

GRV-346

[github-actions]

👋 Thanks for opening this issue! It has been automatically labeled for triage.

[coderabbitai]

🔗 Related PRs

#268 - chore(coderabbit): reconcile config into a single version-controlled source of truth [merged]
#411 - ci: add agentic verification loop workflow and refactor state log [merged]
#693 - fix(ci): unblock the entire PR backlog — resolve invalid agent_network.json + restore apps/web type-check script [merged]
#872 - ci: enforce deterministic agent completion truth gate [merged]
#1142 - fix(security): scope uv.lock checksum false-positive allowlist (#1141) [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!