Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Configure CodeRabbit as a CI gate with GitHub Actions, branch protection, and review enforcement. Use when setting up CodeRabbit as a required check, gating merges on review approval, or integrating CodeRabbit status into your CI pipeline. Trigger with phrases like "coderabbit CI", "coderabbit GitHub Actions", "coderabbit required check", "coderabbit merge gate", "coderabbit CI pipeline".
.claude/skills/jeremylongshore-coderabbit-ci-integration/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 46% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 159% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 108% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 33% | 0% |
Build a merge policy from observable provider state. CodeRabbit review state, GitHub Checks ingestion, and branch protection are separate contracts.
references/official-docs.md and re-check any time-sensitive contract before execution.reviews.request_changes_workflow controls whether CodeRabbit can submit a changes-requested review.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 repository-owner approval before changing branch rules, request-changes behavior, or bypass procedures. Keep analysis and drafts local until approval is explicit, and record who approved the action and its scope.
A merge-policy record, observed check inventory, reviewed patch, pilot evidence, and rollback procedure. 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. |
Replace a guessed coderabbitai required check with observed review and check evidence.
Pilot request-changes behavior while retaining required tests and secret scanning.
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-01 | fail→fail | 16,549 | 8,455 | -49% | 1 | 1 | 0% | 3,131 | 3,586 | +15% | 0 | 0 | — |
case-02 | fail→fail | 15,092 | 10,529 | -30% | 1 | 1 | 0% | 3,063 | 4,168 | +36% | 0 | 0 | — |
case-03 | fail→pass | 13,590 | 9,554 | -30% | 1 | 1 | 0% | 2,551 | 3,725 | +46% | 0 | 0 | — |
case-04 | fail→fail | 17,985 | 17,027 | -5% | 1 | 1 | 0% | 3,330 | 5,040 | +51% | 0 | 0 | — |
case-05 | fail→fail | 15,997 | 13,707 | -14% | 1 | 1 | 0% | 2,684 | 4,322 | +61% | 0 | 0 | — |
case-06 | pass→pass | 14,096 | 13,426 | -5% | 1 | 1 | 0% | 2,329 | 4,218 | +81% | 0 | 0 | — |
case-07 | fail→pass | 4,865 | 2,583 | -47% | 1 | 1 | 0% | 885 | 2,291 | +159% | 0 | 0 | — |
case-08 | fail→pass | 6,300 | 4,022 | -36% | 1 | 1 | 0% | 1,146 | 2,385 | +108% | 0 | 0 | — |
case-09 | pass→pass | 5,624 | 2,437 | -57% | 1 | 1 | 0% | 1,082 | 2,233 | +106% | 0 | 0 | — |
case-10 | pass→pass | 7,031 | 2,947 | -58% | 1 | 1 | 0% | 1,281 | 2,277 | +78% | 0 | 0 | — |
case-11 | pass→pass | 5,054 | 4,074 | -19% | 1 | 1 | 0% | 937 | 2,257 | +141% | 0 | 0 | — |
case-12 | pass→pass | 3,681 | 2,991 | -19% | 1 | 1 | 0% | 787 | 2,401 | +205% | 0 | 0 | — |
case-13 | fail→pass | 9,498 | 1,986 | -79% | 1 | 1 | 0% | 1,580 | 2,088 | +32% | 0 | 0 | — |
case-14 | pass→pass | 6,338 | 1,490 | -76% | 1 | 1 | 0% | 933 | 2,024 | +117% | 0 | 0 | — |
case-15 | pass→pass | 6,590 | 3,581 | -46% | 1 | 1 | 0% | 1,249 | 2,476 | +98% | 0 | 0 | — |
case-16 | pass→pass | 8,811 | 6,279 | -29% | 1 | 1 | 0% | 1,519 | 2,897 | +91% | 0 | 0 | — |
case-17 | pass→pass | 7,690 | 4,500 | -41% | 1 | 1 | 0% | 1,305 | 2,503 | +92% | 0 | 0 | — |
case-18 | pass→pass | 9,968 | 8,406 | -16% | 1 | 1 | 0% | 1,638 | 3,186 | +95% | 0 | 0 | — |
case-19 | fail→pass | 9,798 | 2,534 | -74% | 1 | 1 | 0% | 1,702 | 2,268 | +33% | 0 | 0 | — |
case-20 | pass→pass | 9,133 | 5,991 | -34% | 1 | 1 | 0% | 1,844 | 3,010 | +63% | 0 | 0 | — |
case-21 | pass→pass | 3,544 | 2,093 | -41% | 1 | 1 | 0% | 601 | 2,165 | +260% | 0 | 0 | — |
case-22 | fail→pass | 7,357 | 3,735 | -49% | 1 | 1 | 0% | 1,169 | 2,403 | +106% | 0 | 0 | — |
case-23 | pass→pass | 6,378 | 3,581 | -44% | 1 | 1 | 0% | 1,074 | 2,445 | +128% | 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. 23 cases were attempted. The headline lift of +26 percentage points is the difference between those two pass rates over the 23 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.