audit · raphaelsalaja — Claude Code skill: what it does, how to install
Claude Code skill that audits diffs, branches, or codebases against an entire skill set into a ranked report. Consolidates multi-rule findings before fixes begin.
MIT 176 installs ★ 69 updated 2026-11-14
npx skills add raphaelsalaja/skills --skill audit -a claude-codeWhat audit does
The skill runs code audits across every rule in the raphaelsalaja/skills collection simultaneously. It does not introduce standalone UI heuristics; it directs scope boundaries, enforces check order, and verifies that repository rules in AGENTS.md or CLAUDE.md take priority over default checks. Findings require concrete evidence from running app instances, the diff, or conversation history.
Identified issues reference rule IDs from domain skills like verify, motion, color, copy, and defaults. The workflow consolidates symptoms by root cause: a single broken shared primitive with multiple call sites becomes one item citing multiple locations. Items receive an explicit severity ranking and sequential numbering.
Audit output copies report.html into .audit/<scope>/index.html alongside a generated report.js data file, leaving source folders untouched. The assistant pauses for the operator to select numbered items via interactive picks and only applies fixes for the chosen identifiers. The configuration sets disable-model-invocation: true.
What it helps with
- Run a multi-rule audit across a diff, branch, or entire repository
- Consolidate multiple symptom warnings into unified root-cause issues
- Generate a numbered standalone HTML report outside the source tree
- Apply fixes strictly for user-selected finding numbers
Example prompts for the agent
- Audit the uncommitted diff and open the report
- Run an audit on the current branch against all skills
- Fix items 2 and 4 from the audit report
How to install audit
1Skills CLI (Vercel)
npx skills add raphaelsalaja/skills --skill audit -a claude-code2Manually: put the files in place
~/.claude/skills/audit/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/audit/
1Skills CLI (Vercel)
npx skills add raphaelsalaja/skills --skill audit -a codex2Manually: put the files in place
.agents/skills/audit/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add raphaelsalaja/skills --skill audit -a cursor2Manually: put the files in place
.cursor/skills/audit/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install raphaelsalaja/skills skills/audit2Manually: put the files in place
.github/skills/audit/SKILL.md1Skills CLI (Vercel)
npx skills add raphaelsalaja/skills --skill audit -a gemini-cli2Platform CLI
gemini skills install https://github.com/raphaelsalaja/skills --path skills/auditManually: put the files in place
.devin/skills/audit/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add raphaelsalaja/skills --skill audit -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about audit
How to install audit in Claude Code?
Run npx skills add raphaelsalaja/skills --skill audit -a claude-code. The file is saved to ~/.claude/skills/audit/SKILL.md.
What does the audit skill do?
It orchestrates checks across collection skills, applies local project rules first, merges duplicate findings by root cause, and generates an HTML report in .audit/.
What is required for the audit skill to work?
It relies on the presence of related checks from the raphaelsalaja/skills suite, project instruction files such as CLAUDE.md, and a browser to display report.html.
Source text
SKILL.md · MIT
--- name: audit description: Reviews a change, a session or a whole codebase against every skill in this collection and returns one ranked, numbered report. disable-model-invocation: true --- # Audit This skill reviews work against every rule in this collection at once. It sends each concern to the skill that owns it and merges the results into one ranked report. ## The report, not the rules This skill owns no interface rule. It owns the scope, the order of review, consolidation and the verdict. Each finding cites the rule id from the skill that owns it. A finding needs evidence from the code, the running app or the thread, never taste alone. The report format is in [report.md](report.md). ## 1. State the scope Name what is under review: the uncommitted diff, a branch, a pull request, this session's requests or the whole codebase. An audit with no stated scope is `audit-state-the-scope`. ## 2. Read the project's rules Read `AGENTS.md`, `CLAUDE.md` and the editor rule folders. A written project rule outranks a default here. Flagging code that follows a written rule is `audit-project-rules-first`. ## 3. Run every skill Check the scope against each skill and record what you inspected. | Skill | Check | | --- | --- | | `verify` | Each changed surface opened in the running app | | `scope` | Every request in the thread against the diff | | `reuse` | New components, styles and icons against the shared package | | `restraint` | Unrequested UI, settings, fallbacks, dead code and narrating comments | | `copy` | Every label, description and status string | | `prose` | Each changed post, breakdown or design note, when writing is in scope | | `defaults` | Surfaces, type, sound, focus, names, forms, loading, hotkeys and dialogs | | `motion` | Every transition, animation, gesture, exit and motion blur | | `color` | Text and field contrast in each theme, tokens inside components and the accent | | `gather` | Each reference board: platform, links and the part to take | | `match` | Each named reference, when the work copies one | | `rule` | Repeated corrections with no written rule, and edits that break one | | `viewports` | Each changed layout at every width, when layout is in scope | | `invariants` | Flags read together and unchecked assumptions, when state is in scope | | `ship` | Stray files, secrets, commits and checks, when git is in scope | Skipping a skill without saying why is `audit-run-every-skill`. ## 4. Merge by root cause One cause that breaks rules in two skills is one finding with both ids. Five call sites with the same local fix is one finding with five locations. Listing each symptom on its own line is `audit-one-row-per-cause`. ## 5. Number the fixes and open the report Rank findings by severity, then by reach. Number them so the user can answer "fix 1, 2 and 5". Copy [report.html](report.html) to `.audit/<scope>/index.html`, an ignored folder and never the source tree. Write `report.js` beside it in the shape its header comment shows, then open it. The report's look ships with the skill, so write data only. A page with its own markup or styles is `audit-shipped-report`. The user marks findings with Pick and pastes them back with Copy Picks. Wait for the pick, then fix only the numbers chosen. Fixing everything before the user picks is `audit-fix-what-was-picked`. ## The pair Every check is a pair. The broken line sits beside the line that replaces it. | Rule | Broken | Replacement | | --- | --- | --- | | `audit-state-the-scope` | "Audited the app" | "Audited the uncommitted diff: 14 files" | | `audit-project-rules-first` | A border flagged in a repo whose rules require it | The project rule cited, no finding | | `audit-run-every-skill` | `copy` silently skipped | "`copy`: no strings changed" | | `audit-one-row-per-cause` | Five rows for one missing primitive fix | One row, five locations | | `audit-shipped-report` | A report page written from scratch | `report.html` copied, only `report.js` written | | `audit-fix-what-was-picked` | All 18 findings fixed unasked | The numbers the user chose | ## Reporting Use the severity ladder and format in [report.md](report.md). `Block` when any `HIGH` remains, `Approve` otherwise. Never `Approve` work you did not open or read.
© raphaelsalaja · MIT · raphaelsalaja/skills
Similar skills
All skills for Code review and refactoring →audit by other authors comes first.