Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Diagnose and fix CodeRabbit common errors and configuration issues. Use when CodeRabbit is not reviewing PRs, posting duplicate comments, ignoring configuration, or behaving unexpectedly. Trigger with phrases like "coderabbit error", "fix coderabbit", "coderabbit not working", "debug coderabbit", "coderabbit broken".
.claude/skills/jeremylongshore-coderabbit-common-errors/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 31% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 243% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 148% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 1% | 0% |
Diagnose the first failing boundary instead of relying on fixed wait-time tables. Installation, feature-branch configuration, eligibility, limits, and service health require separate evidence.
references/official-docs.md and re-check any time-sensitive contract before execution..coderabbit.yaml comes from the feature branch under review.base_branches adds other targets.@coderabbitai configuration are contract checks.Treat Git-provider sessions, CodeRabbit web sessions, CLI credentials, and CodeRabbit API keys as separate credentials. Use only an already-approved session or secret-manager reference, never print a secret, and do not place credentials in .coderabbit.yaml, source files, logs, or deliverables.
Require owner approval before changing installation scope, organization settings, or pull-request state. Keep analysis and drafts local until approval is explicit, and record who approved the action and its scope.
A redacted timeline, resolved-config evidence, fault classification, smallest fix, and verification result. Include source dates, unknowns, and the exact boundary between observed fact and recommendation.
| Condition | Response | |---|---| | Current contract is unclear or docs disagree | Stop mutation, cite both sources, and request owner resolution. | | Required access or approval is missing | Produce a draft and evidence plan only. | | Validation or pilot behavior differs from expectation | Restore the prior state and retain the failed evidence. | | Output contains secrets or private code | Stop, quarantine the artifact, redact it, and notify the data owner. |
Explain a skipped release PR whose target pattern is absent.
Correct a runbook that says YAML must first be merged to the base branch.
references/official-docs.md.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-20 | pass→pass | 15,348 | 10,456 | -32% | 1 | 1 | 0% | 2,870 | 4,030 | +40% | 0 | 0 | — |
case-21 | pass→pass | 15,711 | 11,303 | -28% | 1 | 1 | 0% | 3,039 | 4,352 | +43% | 0 | 0 | — |
case-01 | fail→pass | 14,877 | 8,031 | -46% | 1 | 1 | 0% | 2,576 | 3,382 | +31% | 0 | 0 | — |
case-02 | pass→pass | 13,985 | 10,588 | -24% | 1 | 1 | 0% | 2,310 | 3,871 | +68% | 0 | 0 | — |
case-22 | pass→pass | 13,443 | 16,564 | +23% | 1 | 1 | 0% | 2,873 | 5,728 | +99% | 0 | 0 | — |
case-03 | fail→pass | 31,111 | 3,380 | -89% | 1 | 1 | 0% | 2,462 | 2,549 | +4% | 0 | 0 | — |
case-04 | pass→pass | 3,249 | 2,431 | -25% | 1 | 1 | 0% | 621 | 2,403 | +287% | 0 | 0 | — |
case-05 | fail→pass | 3,737 | 2,385 | -36% | 1 | 1 | 0% | 714 | 2,447 | +243% | 0 | 0 | — |
case-06 | pass→pass | 4,204 | 3,306 | -21% | 1 | 1 | 0% | 788 | 2,586 | +228% | 0 | 0 | — |
case-07 | pass→pass | 6,663 | 2,170 | -67% | 1 | 1 | 0% | 1,159 | 2,376 | +105% | 0 | 0 | — |
case-08 | fail→pass | 4,966 | 2,487 | -50% | 1 | 1 | 0% | 970 | 2,410 | +148% | 0 | 0 | — |
case-09 | pass→pass | 3,321 | 1,971 | -41% | 1 | 1 | 0% | 621 | 2,311 | +272% | 0 | 0 | — |
case-10 | pass→pass | 5,578 | 2,328 | -58% | 1 | 1 | 0% | 1,037 | 2,399 | +131% | 0 | 0 | — |
case-11 | fail→pass | 12,373 | 2,257 | -82% | 1 | 1 | 0% | 2,339 | 2,360 | +1% | 0 | 0 | — |
case-12 | pass→pass | 5,242 | 2,487 | -53% | 1 | 1 | 0% | 907 | 2,405 | +165% | 0 | 0 | — |
case-13 | pass→pass | 2,228 | 1,445 | -35% | 1 | 1 | 0% | 352 | 2,206 | +527% | 0 | 0 | — |
case-14 | fail→pass | 10,403 | 4,240 | -59% | 1 | 1 | 0% | 1,635 | 2,625 | +61% | 0 | 0 | — |
case-15 | pass→pass | 5,031 | 2,675 | -47% | 1 | 1 | 0% | 895 | 2,377 | +166% | 0 | 0 | — |
case-16 | pass→pass | 2,849 | 2,686 | -6% | 1 | 1 | 0% | 489 | 2,493 | +410% | 0 | 0 | — |
case-17 | pass→pass | 7,492 | 3,658 | -51% | 1 | 1 | 0% | 1,338 | 2,567 | +92% | 0 | 0 | — |
case-18 | pass→pass | 6,743 | 3,543 | -47% | 1 | 1 | 0% | 1,089 | 2,606 | +139% | 0 | 0 | — |
case-19 | pass→pass | 2,182 | 4,225 | +94% | 1 | 1 | 0% | 332 | 2,465 | +642% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.