skills
CLAUDE.md1 starsChanged 16 days agoMITView on GitHub
This repository is a library of Claude Code skills. If you are editing or adding skills here, follow these conventions. Categories are engineering, productivity, and misc. See CONTEXT.md for what each means and for the shared vocabulary. Every skill is a folder containing a single SKILL.md: - name must equal the folder name. - description is the trigger surface for model-invoked skills. Write it so Claude can tell from the description alone whether the skill applies. Lead with "Use when…"…
In this repository44 other files it gives its agents.
- bigquery-superpowersSkill
- building-data-pipelinesSkill
- data-quality-checksSkill
- dbt-incremental-modelsSkill
- dimensional-modelingSkill
- reviewing-sqlSkill
- writing-dbt-modelsSkill
- arenaSkill
- authority-boundarySkill
- beadflowSkill
- blast-radiusSkill
- control-uiSkill
- create-prSkill
- create-verification-skillSkill
- deslopSkill
- diagnosing-bugsSkill
- feedback-to-infrastructureSkill
- finding-concurrency-bugsSkill
- fixing-and-merging-prsSkill
- foundational-thinkingSkill
- grilling-frontend-prototypingSkill
- improve-harnessSkill
- maintain-verification-skillSkill
- make-operations-idempotentSkill
- metamorphic-testingSkill
- minimize-reader-loadSkill
- preflightSkill
- prototypeSkill
- recon-before-actionSkill
- redesign-from-first-principlesSkill
- resolve-merge-conflictsSkill
- tddSkill
- whySkill
- writing-great-skillsSkill
- catch-me-upSkill
- cross-critiqueSkill
- definition-of-doneSkill
- experience-firstSkill
- grillingSkill
- grill-with-docsSkill
- handoffSkill
- never-block-on-the-humanSkill
- show-me-your-workSkill
- unslopSkill
The file
We split this CLAUDE.md into the rules it gives an agent
4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
Every skill is a folder containing a single SKILL.md:
- Very specific: line 39
One purpose per skill: If a skill is doing two things, split it or make it a composition that chains two smaller skills.
- Very specific: line 41
Write our own prose: When a skill is inspired by someone else's, capture the idea in our own words — do not copy text.
- Very specific: line 46
Keep it composable: Prefer referencing an existing skill (/grilling) over duplicating its body.
≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md 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
CLAUDE.md — Working in this repo
This repository is a library of Claude Code skills. If you are editing or adding skills here, follow these conventions.
Layout
skills/<category>/<skill-name>/SKILL.mdCategories are engineering, productivity, and misc. See CONTEXT.md for
what each means and for the shared vocabulary.
SKILL.md format
Every skill is a folder containing a single SKILL.md:
---
name: skill-name # kebab-case, matches the folder name
description: <one sharp sentence describing WHEN to use this skill>
disable-model-invocation: true # OPTIONAL — only for user-invoked slash commands
---
# Skill Title
<body: the instructions Claude follows when this skill is active>namemust equal the folder name.descriptionis the trigger surface for model-invoked skills. Write it so Claude can tell from the description alone whether the skill applies. Lead with "Use when…" and include concrete trigger phrases where helpful.- Omit
disable-model-invocationfor model-invoked skills (the default).
Authoring rules
- One purpose per skill. If a skill is doing two things, split it or make it a composition that chains two smaller skills.
- Write our own prose. When a skill is inspired by someone else's, capture
the idea in our own words — do not copy text. Add an
_inspired-bynote at the bottom and credit the source in the README Thank Yous. - Record provenance. End each skill with a short note: original, vendored, or inspired-by (with a link).
- Keep it composable. Prefer referencing an existing skill (
/grilling) over duplicating its body.
Adding a skill — checklist
- Create
skills/<category>/<name>/SKILL.mdwith valid frontmatter. namematches the folder;descriptionis a sharp "Use when…" sentence.- Body is in our own voice; provenance note at the bottom.
- Add a row to the catalog table in
README.md. - If it's a port/inspiration from a new source, add a Thank Yous line.
- If it competes with an existing skill over the same prompts, add an
evals/<name>.eval.mdwith should-fire / should-NOT-fire cases. - Run
node scripts/validate-skills.mjs(frontmatter + invocation-dep rule).
Experiments — the wip/ folder
Drafts and experiments live in top-level wip/, not under skills/. That
keeps them out of the live catalog, out of model invocation, and out of CI until
they're ready. Promote with git mv wip/<name> skills/<category>/<name>, then
validate, add a catalog row, and note provenance. See wip/README.md.
Plugin packaging (later)
The repo is structured so it can become an installable Claude Code plugin by
dropping in a .claude-plugin/ manifest. No restructuring of skills/ is
required when that day comes.
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.

