agentsclimarketplace

Diff audit

Skill codecraftersautomation/claude-code/plugins/ccp/skills/diff-audit

Claude Code plugins, MCP servers & skills by CodeCrafters Automation. Includes Change Claude Profile (ccp): hand off / take over work between Claude Code profiles (claude-team ⇄ claude-max) and keep their plugins in sync.

Install
npx -y skills add codecraftersautomation/claude-code --skill diff-audit

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Audit the current git diff before a handoff or commit to catch problems. Use when the user asks to "review my changes", "audit the diff", "check before handoff/commit", "did I leave anything bad in here", or before /ccp:handoff. Looks for accidental edits, generated/committed-by-mistake files, secrets, formatting noise, unfinished TODOs, risky permission changes, and destructive changes. Writes .claude/ccp/handoff/DIFF_AUDIT.md and does NOT modify files unless explicitly asked.

SKILL.md

2.3 KB, 432 tokens by cl100k_base, as published. Nobody here has run it

CCP: diff-audit

Review the working diff with a skeptical eye before it's handed to another profile or committed. The goal is to catch the things that quietly cause problems later: a debug print left in, a secret pasted into a config, a generated file that shouldn't be tracked, a chmod 777, a deletion that wasn't intended.

Steps

  1. Inspect the diff:

    git diff
    git diff --staged
    git status
    
  2. Look for each of these categories and report findings with file + line references:

    • Accidental edits — debug prints, commented-out code, stray whitespace-only churn, leftover scratch changes.
    • Generated / vendored files that probably shouldn't be committed (build output, node_modules, lockfile churn that doesn't belong, large generated assets).
    • Secrets — tokens, API keys, passwords, .env values, private keys. Report the location, not the value.
    • Formatting issues — diffs that are 90% reformatting and obscure the real change; inconsistent indentation.
    • Incomplete TODOsTODO, FIXME, XXX, HACK, half-finished functions.
    • Risky permission changes — files becoming executable, mode 777, ownership changes.
    • Destructive changes — large deletions, removed tests, dropped error handling, schema/migration drops.
  3. Write .claude/ccp/handoff/DIFF_AUDIT.md with a findings list grouped by category, each with severity (info / warning / blocker) and a one-line recommendation.

Read-only by default

Do not modify, stage, revert, or commit anything unless the user explicitly asks you to fix something. The audit informs; the human decides.

What this skill must never do

  • Never switch accounts or touch credentials/OAuth/Keychain data.
  • Never install plugins.
  • Never reprint a discovered secret in full — reference its location.

Gives 0 of the 12 instructions most audit compliance skills give in 432 tokens

Counted across 936 of the 1,487 authors here whose files we hold, read 2026-08-06

  • group findings by severityin 44 of 936
  • Fetch latest guidelines before each reviewin 43 of 936, across 3 files
  • Check files against all fetched rulesin 42 of 936, across 2 files
  • Output findings in terse file:line formatin 41 of 936, across 3 files
  • Ask user which files to review if none specifiedin 41 of 936, across 3 files
  • Read specified files or prompt user for filesin 39 of 936, across 1 file
  • generate the audit reportin 39 of 936, across 36 files
  • assign a severity to every findingin 25 of 936
  • run automated accessibility scansin 23 of 936, across 13 files
  • map findings to WCAG criteriain 20 of 936, across 10 files
  • confirm audit scopein 19 of 936, across 9 files
  • check title tags and meta descriptions for uniquenessin 19 of 936, across 5 files

Said here and by no other author read

  • run git status
  • report accidental edits
  • report generated files
  • report secrets
  • report formatting issues
  • report incomplete TODOs

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.