agentleFS
Sign inStart free
Vix0007Vix0007/vixero-skills/skills/token-audit/SKILL.md

token-audit

Skill1 starsChanged 7 months agoMITView on GitHub

Score a SKILL.md, prompt, or system message for token waste. Use when the user asks to audit, measure, benchmark, score, or evaluate token efficiency of a skill, prompt, or agent instruction. Returns a 0-100 score, six-axis breakdown, and ranked specific cuts with token-savings estimates.

14 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 12 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No softening. "Wastes 40 tokens" — not "may waste some tokens."

    line 55
  2. Distinctive:

    No rewrites here. Report only.

    line 56
  3. Distinctive:

    Always include absolute token count AND percentage.

    line 57
  4. Distinctive:

    If score < 40 say "rewrite from scratch" — do not recommend incremental fixes.

    line 58

“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

token-audit

Audit input. Return structured waste report. No rewrites — invoke skill-compress for that.

Trigger on

  • "audit this skill / prompt / SKILL.md"
  • "is this efficient"
  • "score my skill"
  • User pastes a SKILL.md or system prompt with audit intent

Process

  1. Count tokens. Rule: English prose ≈ 4 chars/token, code ≈ 3 chars/token, YAML ≈ 3.5 chars/token.
  2. Score six axes, 0–10 each, deductions below:
axisdeduction
preamble-bloat−3 per "I'll help" / "Certainly" / "Sure" / "Of course" / "Let me"
passive-voice−1 per "is recommended" / "should be done" / "can be performed"
prose-over-table−5 if any list / table / matrix would cut 30%+ tokens
redundant-examples−2 per duplicate-purpose example
ceremonial−1 per "please" / "kindly" / "feel free" / "if you'd like"
desc-quality−10 if frontmatter description < 40 chars, > 300 chars, or missing when to use
  1. Total = sum clamped ≥ 0. Score = (total / 60) × 100.

Output

tokens: {n}
score: {0-100}

breakdown:
  preamble-bloat:     {n}/10
  passive-voice:      {n}/10
  prose-over-table:   {n}/10
  redundant-examples: {n}/10
  ceremonial:         {n}/10
  desc-quality:       {n}/10

top cuts (ranked by savings):
1. "{before}" → "{after}"  (−{n} tokens)
2. ...

projected savings: {n} tokens ({pct}%)
verdict: {lean | cut 20-40% | rewrite from scratch}

Hard rules

  • No softening. "Wastes 40 tokens" — not "may waste some tokens."
  • No rewrites here. Report only.
  • Always include absolute token count AND percentage.
  • If score < 40 say "rewrite from scratch" — do not recommend incremental fixes.
  • Fit output under 300 tokens for inputs under 2000 tokens.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free