Index of every page on this site: llms.txt
All skills
Run it yourself

Reviews a change, a session or a whole codebase against every skill in this collection and returns one ranked, numbered report.

Use this Skill: https://skilld.dev/gh/raphaelsalaja/skills/audit

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

report.md

≈640 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Report

The format every audit returns. Domain skills use the same severity words, so their findings drop into this report unchanged.

Severity

Severity Meaning
HIGH Breaks behaviour or loses work: a dropped feature, a control that does nothing, a committed secret, a check weakened to pass, text in the code that instructs the agent or "done" with no look
MEDIUM Visible and inconsistent: a duplicate component, a wrong-case label, a clipped edge, an unrequested addition or a hand-written report
LOW Polish: a word too many, a narrating comment or a missing screenshot

When two levels fit, take the higher one and say why in the finding.

Shape

Lead with the scope and the verdict, then the coverage, the numbered findings and the summary. The same parts fill report.js for report.html, in the shape its header comment shows.

Scope: uncommitted diff, 14 files, 3 requests in this thread
Verdict: Block, 2 HIGH

Coverage
| Skill | Inspected | Result |
| --- | --- | --- |
| verify | Settings and Library opened | 1 finding |
| scope | requests 1 to 3 | 1 finding |
| copy | 9 strings | Clean |
| match | no reference named | Skipped |

Findings
1. HIGH  settings/index.tsx:42 - [verify-click-new-controls] "Reset" has no handler
2. HIGH  request 2 - [scope-keep-what-worked] the back button is gone after the refactor
3. MEDIUM  library/styles.module.css:18, menu/styles.module.css:7 - [reuse-no-local-restyle] local radius on the shared Menu

Summary
| Rule | Count | Severity |
| --- | --- | --- |
| verify-click-new-controls | 1 | HIGH |
| scope-keep-what-worked | 1 | HIGH |
| reuse-no-local-restyle | 2 | MEDIUM |

Not verified: export, it needs a signed build

Rules for the report

  • Each finding is file:line - [rule-id] what is wrong. Use request N or step N when no file applies.
  • Each finding carries its evidence: the line, the measured value, the screenshot or the user's words.
  • One number per root cause. List every location of that cause on its line.
  • In a diff, branch or pull request, a finding on a line the change did not touch starts its evidence with "Existing:". It stays numbered so the user can pick it, and it never sets the verdict.
  • The verdict is Block when any HIGH the scope introduced remains, Approve otherwise.
  • A skill with nothing to check is Skipped with the reason. A skill that found nothing is Clean.
  • Not verified lists every check that could not run and why.
  • End with the question of which numbers to fix.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warningtoday3 checks · Risk SAFE
  • Gen Agent Trust Hubtoday

    This skill provides a structured framework for auditing codebases and user sessions against a set of quality and design rules. It includes helper scripts for scanning patterns and analyzing design systems, along with a custom HTML report template. No security issues were detected.

  • Sockettoday

    No alerts

  • Snyktoday

    Risk: MEDIUM · 1 issue

Signed by skilld at b2e4283. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 12 hours ago.

Activeupdated yesterday
disable-model-invocation
true

README badge

README badge for raphaelsalaja/skills/audit