Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Reviews pull requests for compliance regressions. Scans code diffs for security and compliance violations, flags issues, and suggests fixes aligned with frameworks like SOC 2, ISO 27001, NIST 800-53.
.claude/skills/grcengclub-audit-ready-pr-reviewer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -34% | 0% |
Reviews GitHub/GitLab pull requests specifically for compliance regressions. Shifts compliance "left" into the developer's daily workflow.
Review a PR for SOC 2 compliance:
bashnode plugins/grc-engineer/scripts/review-pr.js myorg/infrastructure 42 SOC2
Review a PR for ISO 27001:
bashnode plugins/grc-engineer/scripts/review-pr.js myorg/infrastructure 42 ISO27001
Review a PR with custom framework:
bashnode plugins/grc-engineer/scripts/review-pr.js myorg/infrastructure 42 NIST80053
Posts GitHub comments with:
markdown⚠️ **Compliance Warning: SOC 2 CC6.1 - Least Privilege** This PR introduces an IAM role with `AdministratorAccess`, which violates the Least Privilege principle. **Issue:** Line 23 in `terraform/iam.tf` assigns full administrative access. **Suggested Fix:**
resource "aws_iam_role" "app_role" { assume_role_policy = jsonencode({ Version = "2012-10-17" Statement = { Effect = "Allow" Action = "s3:GetObject", "s3:PutObject" ] Resource = "arn:aws:s3:::my-bucket/" }] }) }
**Control Reference:** SOC 2 CC6.1, NIST 800-53 AC-6
GITHUB_TOKEN environment variable (requires repo scope)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-13 | pass→pass | 8,568 | 3,633 | -58% | 1 | 1 | 0% | 1,309 | 885 | -32% | 0 | 0 | — |
case-01 | fail→pass | 9,647 | 8,164 | -15% | 1 | 1 | 0% | 1,630 | 1,759 | +8% | 0 | 0 | — |
case-02 | fail→pass | 22,426 | 14,955 | -33% | 1 | 1 | 0% | 4,017 | 3,284 | -18% | 0 | 0 | — |
case-03 | fail→pass | 19,100 | 16,796 | -12% | 1 | 1 | 0% | 3,199 | 3,675 | +15% | 0 | 0 | — |
case-04 | pass→pass | 15,767 | 12,266 | -22% | 1 | 1 | 0% | 2,486 | 2,381 | -4% | 0 | 0 | — |
case-05 | pass→pass | 14,629 | 12,869 | -12% | 1 | 1 | 0% | 3,157 | 3,917 | +24% | 0 | 0 | — |
case-06 | pass→pass | 7,091 | 5,255 | -26% | 1 | 1 | 0% | 1,252 | 1,471 | +17% | 0 | 0 | — |
case-07 | fail→fail | 9,410 | 3,154 | -66% | 1 | 1 | 0% | 964 | 1,009 | +5% | 0 | 0 | — |
case-08 | pass→pass | 8,044 | 1,992 | -75% | 1 | 1 | 0% | 1,156 | 873 | -24% | 0 | 0 | — |
case-09 | fail→pass | 6,114 | 1,469 | -76% | 1 | 1 | 0% | 880 | 782 | -11% | 0 | 0 | — |
case-10 | fail→pass | 8,663 | 1,870 | -78% | 1 | 1 | 0% | 1,245 | 822 | -34% | 0 | 0 | — |
case-11 | fail→pass | 10,726 | 3,162 | -71% | 1 | 1 | 0% | 1,539 | 1,092 | -29% | 0 | 0 | — |
case-12 | pass→pass | 6,299 | 2,457 | -61% | 1 | 1 | 0% | 907 | 954 | +5% | 0 | 0 | — |
case-14 | pass→pass | 9,088 | 3,937 | -57% | 1 | 1 | 0% | 1,418 | 1,190 | -16% | 0 | 0 | — |
case-15 | pass→pass | 8,929 | 3,689 | -59% | 1 | 1 | 0% | 1,368 | 1,117 | -18% | 0 | 0 | — |
case-16 | pass→pass | 7,350 | 2,242 | -69% | 1 | 1 | 0% | 1,052 | 882 | -16% | 0 | 0 | — |
case-17 | pass→pass | 8,035 | 2,950 | -63% | 1 | 1 | 0% | 1,260 | 1,009 | -20% | 0 | 0 | — |
case-18 | pass→pass | 8,116 | 2,827 | -65% | 1 | 1 | 0% | 1,352 | 933 | -31% | 0 | 0 | — |
case-19 | fail→pass | 10,879 | 1,685 | -85% | 1 | 1 | 0% | 1,677 | 825 | -51% | 0 | 0 | — |
case-20 | pass→pass | 15,887 | 3,333 | -79% | 1 | 1 | 0% | 2,365 | 1,097 | -54% | 0 | 0 | — |
case-21 | pass→pass | 14,218 | 5,649 | -60% | 1 | 1 | 0% | 2,146 | 1,481 | -31% | 0 | 0 | — |
case-22 | fail→pass | 6,185 | 3,342 | -46% | 1 | 1 | 0% | 1,053 | 1,066 | +1% | 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 +36 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.