lean-audit
Skill4 starsChanged 2 months agoMITView on GitHub
Use when the user asks to audit a whole repository for over-engineering, code bloat, unnecessary dependencies, speculative abstractions, or what can be deleted or simplified.
3 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- asta-nguyen/kamehadbedited, first seen Oct 16, 2026
- FareedKhan-dev/lean-agenttheir own version, first seen Oct 13, 2026
- utk2103/Prompt-Studiotheir own version, first seen Aug 22, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- asta-nguyen/kamehadb★ 14100% same
Wrote their own version · 2
- utk2103/Prompt-Studio★ 8same name
- FareedKhan-dev/lean-agent★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository15 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- writing-skillsSkill
- brainstorm-featureSkill
- context-handoffSkill
- document-wikiSkill
- estimate-featureSkill
- implement-taskSkill
- plan-featureSkill
- read-codebase-contextSkill
- review-and-verifySkill
- setup-codebaseSkill
- setup-openezSkill
- systematic-debuggingSkill
- using-devkitSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Run a read-only whole-repo simplicity audit.
- Distinctive: line 15
Use for audits of over-engineering, bloat, or simplification.
- Distinctive: line 16
Do not use for diff-only review, correctness, security, performance, generic test coverage, or style.
- Very specific: line 25
If a required area or repository-wide search is unavailable, report Audit incomplete: <specific limitation> and do not make a lean verdict.
- Very specific: line 51
Every finding needs an exact repository-relative path:line, evidence, and a replacement.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Lean Audit
Overview
Run a read-only whole-repo simplicity audit. Report validated cuts; do not edit, delete, commit, open a PR, or post comments.
When to use
Use for audits of over-engineering, bloat, or simplification. Do not use for diff-only review, correctness, security, performance, generic test coverage, or style; route those concerns to review or debugging.
Process
- Establish root. Read applicable instructions and inspect
git statuswithout changing it. Record root, exclusions (.git, generated output, vendored code,node_modules, and build caches), and every inspected first-party area. If the user explicitly includes an excluded path, state that override. If a required area or repository-wide search is unavailable, reportAudit incomplete: <specific limitation>and do not make a lean verdict. - Inventory source, structure, entrypoints, manifests, dependencies,
configuration, commands, and tests. Use
rgand direct reads. Before judging a candidate, use FFF multi-pattern grep (fallback:rg) to search callers, registrations, and dynamic-reference variants, then inspect relevant configuration, tests, and extension requirements. - Report only validated cuts with one of these tags:
delete,stdlib,native,yagni, orshrink. Omit ambiguous candidates. Preserve explicit requirements, repository conventions, compatibility boundaries, required validation, and behavior-protecting tests. Route out-of-scope concerns to review/debugging. - Rank the largest credible cut first. Report savings only when supported; never imply that a change was applied.
Output
# Lean audit
Scope: <repository root>; exclusions: <paths>; inspected: <first-party areas>
1. delete: <path:line> <cut>. <replacement>. evidence: <paths/searches>
net: -<N> lines, -<M> dependencies possible.Every finding needs an exact repository-relative path:line, evidence, and a
replacement. For shrink, show the shorter equivalent form. After complete
coverage with no validated cut, keep the header and scope, then output
Lean already. Ship. and omit findings and net. If coverage is incomplete,
output the limitation instead. End with the Next: block from using-devkit
(Next: none after a complete report).
Quick reference
| Tag | Meaning |
delete | Remove code or a dependency with no required use. |
stdlib | Replace custom behavior with a standard-library API. |
native | Replace custom machinery with a platform-native feature. |
yagni | Remove speculative or unused abstraction. |
shrink | Keep behavior while making code materially smaller. |
Example
src/format.js:18 — stdlib: replace the hand-written extension parser with
node:path.extname; evidence: callers and tests cover the same input boundary.
Common mistakes
- Calling a style preference bloat without behavior or reuse evidence.
- Calling code dead after checking only one direct occurrence.
- Removing a factory or validation despite an explicit extension contract.
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
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.

