Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Security vulnerability expert covering OWASP Top 10 and common security issues. Use when conducting security audits or reviewing code for vulnerabilities.
.claude/skills/security-auditor/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 4% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 16% | 0% |
Expert in identifying security vulnerabilities following OWASP Top 10 and security best practices.
Activates when you:
Use the current OWASP Top 10 as an awareness taxonomy, not as proof of complete security coverage. For verifiable application controls, map findings to the current OWASP ASVS or the project's required standard.
Run security audit:
bashpython3 scripts/security_audit.py --name <service-name> --output security-audit.md
Check for secrets:
bashpython3 scripts/find_secrets.py .
references/owasp.md - OWASP Top 10 detailsreferences/checklist.md - Security audit checklistreferences/remediation.md - Vulnerability remediation guide| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 13,221 | 29,927 | +126% | 1 | 1 | 0% | 1,344 | 1,929 | +44% | 0 | 0 | — |
case-02 | fail→fail | 20,347 | 18,891 | -7% | 1 | 1 | 0% | 2,332 | 1,307 | -44% | 0 | 0 | — |
case-03 | fail→fail | 23,102 | 18,130 | -22% | 1 | 1 | 0% | 2,442 | 2,214 | -9% | 0 | 0 | — |
case-04 | fail→pass | 18,669 | 21,124 | +13% | 1 | 1 | 0% | 2,217 | 3,130 | +41% | 0 | 0 | — |
case-05 | pass→pass | 29,190 | 24,918 | -15% | 1 | 1 | 0% | 3,504 | 3,716 | +6% | 0 | 0 | — |
case-06 | fail→pass | 15,155 | 7,305 | -52% | 1 | 1 | 0% | 1,652 | 1,132 | -31% | 0 | 0 | — |
case-07 | fail→pass | 12,302 | 14,689 | +19% | 1 | 1 | 0% | 1,148 | 1,196 | +4% | 0 | 0 | — |
case-08 | fail→pass | 15,125 | 15,963 | +6% | 1 | 1 | 0% | 1,776 | 2,463 | +39% | 0 | 0 | — |
case-09 | fail→pass | 10,706 | 12,755 | +19% | 1 | 1 | 0% | 1,863 | 2,155 | +16% | 0 | 0 | — |
case-10 | fail→pass | 5,420 | 54,588 | +907% | 1 | 1 | 0% | 992 | 1,746 | +76% | 0 | 0 | — |
case-11 | fail→pass | 3,571 | 9,971 | +179% | 1 | 1 | 0% | 688 | 1,394 | +103% | 0 | 0 | — |
case-12 | pass→pass | 12,567 | 12,537 | -0% | 1 | 1 | 0% | 1,276 | 2,047 | +60% | 0 | 0 | — |
case-13 | pass→pass | 6,104 | 14,310 | +134% | 1 | 1 | 0% | 817 | 1,712 | +110% | 0 | 0 | — |
case-14 | pass→pass | 11,462 | 6,423 | -44% | 1 | 1 | 0% | 946 | 1,720 | +82% | 0 | 0 | — |
case-15 | pass→pass | 6,146 | 13,478 | +119% | 1 | 1 | 0% | 901 | 1,775 | +97% | 0 | 0 | — |
case-16 | fail→pass | 3,034 | 2,193 | -28% | 1 | 1 | 0% | 389 | 1,038 | +167% | 0 | 0 | — |
case-17 | pass→pass | 18,329 | 9,166 | -50% | 1 | 1 | 0% | 1,545 | 1,361 | -12% | 0 | 0 | — |
case-18 | pass→pass | 6,821 | 4,075 | -40% | 1 | 1 | 0% | 1,206 | 1,528 | +27% | 0 | 0 | — |
case-19 | fail→pass | 17,203 | 6,953 | -60% | 1 | 1 | 0% | 2,079 | 1,059 | -49% | 0 | 0 | — |
case-20 | fail→pass | 3,763 | 6,456 | +72% | 1 | 1 | 0% | 434 | 1,054 | +143% | 0 | 0 | — |
case-21 | pass→pass | 18,056 | 18,280 | +1% | 1 | 1 | 0% | 2,794 | 3,383 | +21% | 0 | 0 | — |
case-22 | pass→pass | 14,160 | 16,622 | +17% | 1 | 1 | 0% | 1,858 | 3,207 | +73% | 0 | 0 | — |
case-23 | pass→pass | 9,777 | 13,532 | +38% | 1 | 1 | 0% | 1,746 | 2,648 | +52% | 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 +43 percentage points is the difference between those two pass rates over the 23 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/2/2026 | +23% |
| gemini-3.6-flash | verified | 7/26/2026 | +18% |
Other measured skills in the registry, with their headline benchmark lift.