readme
Skill2 starsChanged 4 months agoMITView on GitHub
Generate or review a README following state-of-the-art practices. Use when creating a README from scratch, improving or restructuring an existing README, choosing badges, or auditing a README against best-practice checklists.
38 other projects use a version of this.
Adopted by
Most recently taken up by
- GoBeromsu/craft-skillstheir own version, first seen Nov 16, 2026
- gabrielSantosLima/harnesstheir own version, first seen Nov 16, 2026
- wufangci/READMEtheir own version, first seen Nov 16, 2026
- yassine-krichen/claude-code-agentic-setuptheir own version, first seen Nov 16, 2026
- computerlovetech/skillstheir own version, first seen Nov 16, 2026
- Xia-Ataraxia/craft-skillstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use favicon, build and 6 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
cpiprintSkillfavicon
Generate favicons from a source image
Also in 3 projects that use this
Microsoft AzureSkillbuild
Build the TypeSpec Rust emitter from source. Run this after cloning the repo, after pulling changes, or when TypeScript source files have been modified.
Also in 3 projects that use this
GitHubSkillrefactor
Surgical code refactoring to improve maintainability without changing behavior. Covers extracting functions, renaming variables, breaking down god functions, improving type safety, eliminating code smells, and applying design patterns. Less drastic than repo-rebuilder; use for gradual improvements.
Also in 3 projects that use this
BlockSkillplan
Draft a TDD, PRD, or lightweight plan in the conversation without committing it.
Also in 3 projects that use this
GoogleSkillprototype
Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.
Also in 3 projects that use this
SentrySkillcommit
Create commit messages following Sentry conventions. Use when committing code changes, writing commit messages, or formatting git history. Follows conventional commits with Sentry-specific issue references.
Also in 3 projects that use this
Other versions38 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 38
- getkyo/kyo★ 820same name
- statelyai/skills★ 50same name
- sampleXbro/agentsmesh★ 27same name
- wpgaurav/WordPress-skills★ 19same name
- danvega/skills★ 15same name
- kasperjunge/agent-resources-legacy★ 9same name
- harlan-zw/brundlefly★ 7same name
- lawless-m/claude-skills★ 6same name
- ephemerallyCaden/write-readme★ 5same name
- computerlovetech/skills★ 4same name
28 more not listed
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 repository4 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- refresh-researchSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 5 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 38 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
If the arguments don't say, infer from context: an existing non-trivial README means review.
- Very specific: line 24
Read ${CLAUDE_SKILL_DIR}/references/project-types.md, then inspect the repo (manifests, directory layout, entry points) and pick ONE …
- Very specific: line 38
Read ${CLAUDE_SKILL_DIR}/preferences/about.md, then look for a profile in this order and use the first found: a path named in the …
- Very specific: line 53
Never invent facts: Every command, feature, version, and badge must be verified against the actual codebase (run --help, read the manifest, …
- Distinctive: line 58
Write the one-liner before everything else.
- Distinctive: line 77
If a style profile is active (Step 1b), add a separate Style profile adherence section after the score: each profile rule marked met / not …
- Very specific: line 84
Read ${CLAUDE_SKILL_DIR}/references/ai-readers.md and apply it in both modes: READMEs are now executed by coding agents as much as read by …
- Very specific: line 94
generate / fixing: after writing the one-liner, set the repo description to match, shaped by the active profile's About rule (Step 1b.
- Distinctive: line 97
If not, tell the user the exact text to paste into About.
- Distinctive: line 107
generate: write README.md at the repo root (or the path given), sync the About description (Step 4), then summarize which sections you …
“Common” is measured: kept by at least 6 in 10 of the 38 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.
Files
README standard
You generate or review READMEs against a researched, cited knowledge base. All
substance lives in references/ and templates/ — read the files listed for
your mode; do not work from memory of "typical" READMEs.
Step 0 — Pick a mode
- generate — no README exists, or the user wants one written/rewritten from scratch.
- review — a README exists and the user wants it audited or improved.
If the arguments don't say, infer from context: an existing non-trivial README means review; absent or placeholder README means generate. Improving a weak existing README = review first, then apply fixes.
Step 1 — Classify the project
Read ${CLAUDE_SKILL_DIR}/references/project-types.md, then inspect the repo
(manifests, directory layout, entry points) and pick ONE primary type:
| Type | Template |
| Library / package | templates/library.md |
| CLI tool | templates/cli-tool.md |
| Web app / self-hosted app | templates/web-app.md |
| API / service / framework | templates/api-service.md |
| Research / data project | templates/research-data.md |
| Tiny project / not sure | templates/minimal.md |
Step 1b — Load the active style profile
Read ${CLAUDE_SKILL_DIR}/preferences/about.md, then look for a profile in this
order and use the first found: a path named in the invocation → the target
project's .claude/readme-preferences.md → ~/.claude/readme-preferences.md →
none (use the standard's neutral voice). A profile shapes voice and formatting
only; the standard's correctness rules below always win.
Step 2a — Generate
Read, in order:
${CLAUDE_SKILL_DIR}/references/anatomy.md— section order and which sections this project type requires${CLAUDE_SKILL_DIR}/templates/<type>.md— the starting skeleton${CLAUDE_SKILL_DIR}/references/section-guide.md— per-section writing rules${CLAUDE_SKILL_DIR}/references/badges.md— which badges this project earns, ordering, syntax
Hard rules:
- Never invent facts. Every command, feature, version, and badge must be
verified against the actual codebase (run
--help, read the manifest, check CI config). A README that lies is worse than no README. - Omit sections the project can't fill honestly (no "Roadmap: TBD" stubs).
- Skip badges that would be red or "unknown" on day one.
- Write the one-liner before everything else; if you can't state what the project does and for whom in one sentence, ask the user.
Then shape the prose to the active style profile (Step 1b): correctness first, taste second. If a profile rule would drop required content, keep the content and restructure to honor the profile's spirit.
Step 2b — Review
Read:
${CLAUDE_SKILL_DIR}/references/checklist.md— the scored rubric${CLAUDE_SKILL_DIR}/references/anti-patterns.md— named failure modes
Produce a report with: overall score per the checklist's scheme, the top 3–5 highest-impact fixes first (ranked by reader impact, not section order), each finding tied to a checklist item or named anti-pattern, and concrete replacement text for the worst offender. Verify claims in the README against the repo — stale install commands and dead links are findings, not style notes.
If a style profile is active (Step 1b), add a separate Style profile adherence section after the score: each profile rule marked met / not met, with a concrete rewrite for the biggest miss. Keep it out of the numeric score — taste is not sourced, so it never moves the objective rubric.
Step 3 — Always apply agent-era rules
Read ${CLAUDE_SKILL_DIR}/references/ai-readers.md and apply it in both modes:
READMEs are now executed by coding agents as much as read by humans. Key
implications: copy-pasteable commands, no agent instructions inlined in the
README (link to AGENTS.md if present), non-obvious constraints stated explicitly.
Step 4 — Sync the repo "About"
The README one-liner and the GitHub About description are the same sentence in two places (Identity: they must match). The README file is only half the job.
- generate / fixing: after writing the one-liner, set the repo description to
match, shaped by the active profile's About rule (Step 1b; defaults to the
tagline rule). If
ghis available:gh repo edit <owner/repo> --description "<one-liner>". If not, tell the user the exact text to paste into About. When the repo has a live site (GitHub Pages, deployed app), offer to set the homepage URL and a few topics too — but only the description is required. - review: fetch the current description (
gh repo view <owner/repo> --json description) and compare it to the README one-liner. A mismatch (or an empty About) is an Identity finding, not a nice-to-have. Local clones don't expose this — you must query the repo.
Output
- generate: write
README.mdat the repo root (or the path given), sync the About description (Step 4), then summarize which sections you included, which you omitted and why, and the description you set. - review: present the report, including any README↔About mismatch; only edit the README or the description if the user asked for fixes, not just an audit.
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.

