Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Prowler documentation style guide and writing standards. Trigger: When writing documentation for Prowler features, tutorials, or guides.
.claude/skills/foryourhealth111-pixel-prowler-docs/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 61% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -7% | 0% |
Use this skill when writing Prowler documentation for:
Use Title Case for all headers:
Use when information can be logically divided:
markdownProwler CLI includes: * **Industry standards:** CIS, NIST 800, NIST CSF * **Regulatory compliance:** RBI, FedRAMP, PCI-DSS * **Privacy frameworks:** GDPR, HIPAA, FFIEC
Place keywords at the beginning:
mdximport { VersionBadge } from "/snippets/version-badge.mdx" ## New Feature Name <VersionBadge version="4.5.0" /> Description of the feature...
mdx<Warning> Disabling encryption may expose sensitive data to unauthorized access. </Warning> <Danger> Running this command will **permanently delete all data**. </Danger>
Reference without articles:
docs/
├── getting-started/
├── tutorials/
├── providers/
│ ├── aws/
│ ├── azure/
│ ├── gcp/
│ └── ...
├── api/
├── sdk/
├── compliance/
└── developer-guide/| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,127 | 13,483 | -16% | 1 | 1 | 0% | 2,912 | 3,429 | +18% | 0 | 0 | — |
case-02 | fail→pass | 10,955 | 8,136 | -26% | 1 | 1 | 0% | 1,909 | 2,036 | +7% | 0 | 0 | — |
case-03 | fail→fail | 12,437 | 12,635 | +2% | 1 | 1 | 0% | 2,282 | 3,038 | +33% | 0 | 0 | — |
case-09 | pass→pass | 9,498 | 5,865 | -38% | 1 | 1 | 0% | 1,658 | 1,603 | -3% | 0 | 0 | — |
case-04 | fail→pass | 72,560 | 5,712 | -92% | 1 | 1 | 0% | 1,016 | 1,633 | +61% | 0 | 0 | — |
case-05 | fail→pass | 19,324 | 12,054 | -38% | 1 | 1 | 0% | 3,466 | 2,801 | -19% | 0 | 0 | — |
case-06 | fail→fail | 19,531 | 13,130 | -33% | 1 | 1 | 0% | 3,355 | 2,828 | -16% | 0 | 0 | — |
case-07 | pass→pass | 7,948 | 6,962 | -12% | 1 | 1 | 0% | 1,309 | 1,827 | +40% | 0 | 0 | — |
case-08 | fail→pass | 13,919 | 9,608 | -31% | 1 | 1 | 0% | 2,668 | 2,490 | -7% | 0 | 0 | — |
case-10 | pass→pass | 7,994 | 5,792 | -28% | 1 | 1 | 0% | 1,617 | 1,797 | +11% | 0 | 0 | — |
case-11 | pass→pass | 13,152 | 10,774 | -18% | 1 | 1 | 0% | 2,245 | 2,736 | +22% | 0 | 0 | — |
case-12 | fail→fail | 17,134 | 13,693 | -20% | 1 | 1 | 0% | 2,907 | 3,456 | +19% | 0 | 0 | — |
case-13 | pass→pass | 10,072 | 8,729 | -13% | 1 | 1 | 0% | 1,625 | 2,293 | +41% | 0 | 0 | — |
case-14 | fail→pass | 9,633 | 4,884 | -49% | 1 | 1 | 0% | 1,623 | 1,539 | -5% | 0 | 0 | — |
case-15 | fail→pass | 10,418 | 8,975 | -14% | 1 | 1 | 0% | 1,562 | 2,222 | +42% | 0 | 0 | — |
case-16 | pass→pass | 13,747 | 10,991 | -20% | 1 | 1 | 0% | 2,167 | 2,550 | +18% | 0 | 0 | — |
case-17 | fail→pass | 11,432 | 7,625 | -33% | 1 | 1 | 0% | 1,936 | 2,091 | +8% | 0 | 0 | — |
case-18 | pass→pass | 13,137 | 8,294 | -37% | 1 | 1 | 0% | 2,242 | 2,221 | -1% | 0 | 0 | — |
case-19 | fail→pass | 5,935 | 9,865 | +66% | 1 | 1 | 0% | 941 | 1,525 | +62% | 0 | 0 | — |
case-20 | pass→pass | 9,934 | 8,082 | -19% | 1 | 1 | 0% | 1,975 | 2,255 | +14% | 0 | 0 | — |
case-21 | pass→pass | 9,195 | 7,932 | -14% | 1 | 1 | 0% | 1,771 | 2,310 | +30% | 0 | 0 | — |
case-22 | pass→pass | 4,912 | 4,804 | -2% | 1 | 1 | 0% | 770 | 1,500 | +95% | 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 +41 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.