Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Configure CodeRabbit enterprise access control, seat management, and organization policies. Use when managing who gets AI reviews, configuring organization-level defaults, or implementing access policies for CodeRabbit across teams. Trigger with phrases like "coderabbit SSO", "coderabbit RBAC", "coderabbit enterprise", "coderabbit roles", "coderabbit permissions", "coderabbit seats".
.claude/skills/jeremylongshore-coderabbit-enterprise-rbac/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 176% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 148% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 64% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 57% | 0% |
Use CodeRabbit native administrative roles instead of inferring all access from the Git provider. Map developer seats and administrative authority separately.
references/official-docs.md and re-check any time-sensitive contract before execution.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 an authorized CodeRabbit Admin and security owner before role, seat, default-role, or key changes. Keep analysis and drafts local until approval is explicit, and record who approved the action and its scope.
An access matrix, least-privilege target, exceptions, approved change set, and recertification schedule. 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. |
Separate billing duties using Billing Admin.
Create an Enterprise read-only audit custom role.
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→pass | 19,138 | 15,328 | -20% | 1 | 1 | 0% | 3,465 | 4,736 | +37% | 0 | 0 | — |
case-02 | fail→fail | 17,548 | 14,329 | -18% | 1 | 1 | 0% | 3,440 | 4,537 | +32% | 0 | 0 | — |
case-03 | pass→pass | 13,235 | 9,693 | -27% | 1 | 1 | 0% | 2,193 | 3,467 | +58% | 0 | 0 | — |
case-12 | pass→pass | 6,939 | 4,654 | -33% | 1 | 1 | 0% | 1,229 | 2,590 | +111% | 0 | 0 | — |
case-04 | fail→pass | 5,690 | 4,134 | -27% | 1 | 1 | 0% | 921 | 2,544 | +176% | 0 | 0 | — |
case-05 | pass→pass | 10,291 | 5,858 | -43% | 1 | 1 | 0% | 1,800 | 2,779 | +54% | 0 | 0 | — |
case-06 | pass→pass | 6,744 | 4,297 | -36% | 1 | 1 | 0% | 1,162 | 2,607 | +124% | 0 | 0 | — |
case-07 | fail→pass | 5,147 | 3,390 | -34% | 1 | 1 | 0% | 988 | 2,451 | +148% | 0 | 0 | — |
case-17 | pass→pass | 9,627 | 6,002 | -38% | 1 | 1 | 0% | 1,696 | 2,809 | +66% | 0 | 0 | — |
case-08 | pass→pass | 6,498 | 7,716 | +19% | 1 | 1 | 0% | 1,111 | 3,198 | +188% | 0 | 0 | — |
case-09 | pass→pass | 11,865 | 10,930 | -8% | 1 | 1 | 0% | 1,917 | 3,688 | +92% | 0 | 0 | — |
case-10 | pass→pass | 11,565 | 6,447 | -44% | 1 | 1 | 0% | 1,970 | 2,828 | +44% | 0 | 0 | — |
case-11 | pass→pass | 8,179 | 5,698 | -30% | 1 | 1 | 0% | 1,512 | 2,789 | +84% | 0 | 0 | — |
case-22 | pass→pass | 4,683 | 3,894 | -17% | 1 | 1 | 0% | 1,004 | 2,675 | +166% | 0 | 0 | — |
case-13 | pass→pass | 7,863 | 3,499 | -56% | 1 | 1 | 0% | 1,283 | 2,447 | +91% | 0 | 0 | — |
case-14 | pass→pass | 5,082 | 4,638 | -9% | 1 | 1 | 0% | 861 | 2,797 | +225% | 0 | 0 | — |
case-15 | fail→pass | 11,459 | 7,924 | -31% | 1 | 1 | 0% | 1,909 | 3,139 | +64% | 0 | 0 | — |
case-16 | pass→pass | 11,335 | 6,712 | -41% | 1 | 1 | 0% | 2,024 | 2,926 | +45% | 0 | 0 | — |
case-18 | pass→pass | 9,798 | 3,078 | -69% | 1 | 1 | 0% | 1,628 | 2,256 | +39% | 0 | 0 | — |
case-19 | pass→pass | 7,865 | 3,918 | -50% | 1 | 1 | 0% | 1,421 | 2,469 | +74% | 0 | 0 | — |
case-20 | fail→pass | 8,948 | 3,651 | -59% | 1 | 1 | 0% | 1,531 | 2,406 | +57% | 0 | 0 | — |
case-21 | pass→pass | 10,288 | 6,943 | -33% | 1 | 1 | 0% | 1,930 | 3,388 | +76% | 0 | 0 | — |
case-23 | pass→pass | 9,398 | 9,289 | -1% | 1 | 1 | 0% | 1,889 | 3,523 | +87% | 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 +22 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.