init2
Skill0 starsChanged 56 days agoUnlicenseView on GitHub
Generate a world-class project-specific CLAUDE.md by deeply analyzing this repository.
In this repository9 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- commitSkill
- context-jokeSkill
- gh-pr-consider-commentsSkill
- gh-pr-reviewSkill
- grill-meSkill
- grill-projSkill
- knowledge-interviewSkill
- pull-requestSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
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
- Distinctive: line 26
Create a CLAUDE.md at the project root with this structure:
- Distinctive: line 79
Do NOT generate a thin, generic CLAUDE.md.
≈ “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
Generate a world-class project-specific CLAUDE.md by deeply analyzing this repository. The goal is to capture derived knowledge that saves Claude from re-reading the entire project at the start of every session.
Reading Order (highest signal first)
Work through these in order, reading the most information-dense sources first:
- README.md, CONTRIBUTING.md — intent, setup instructions, conventions
- docs/adrs/, docs/decisions/ — architectural decisions and their context
- docs/* — domain knowledge, architecture guides, API docs
- Package files (package.json, Cargo.toml, go.mod, requirements.txt, etc.) — stack, scripts, dependencies
- Config files (tsconfig, eslint/biome, prettier, terraform, docker, etc.) — coding standards already enforced by tooling
- CI/CD config (.github/workflows/, Jenkinsfile, etc.) — what's tested, how it deploys
- Directory structure (tree, 2 levels deep) — module layout, naming
- Key entrypoints (main/index files, route definitions, handler files) — architecture in practice
- Test setup (test config, a few example test files) — testing patterns, framework, structure, conventions
Output
Create a CLAUDE.md at the project root with this structure:
# Project: <n>
## Overview
<What this project does, its purpose, key domain concepts. 2-3 sentences.>
## Stack
<Detected tech stack with versions where relevant. Be specific.>
## Key Commands
- **Install:** `<command>`
- **Build:** `<command>`
- **Test:** `<command>` (framework: <n>, location: <pattern>)
- **Lint:** `<command>`
- **Dev:** `<command>`
- **Deploy:** `<command or process description>`
## Architecture
<High-level description of how the project is structured — modules, layers, data flow. Derived from directory structure
and entrypoint analysis. Keep it concise but complete enough that someone (or Claude) can orient quickly.>
## Conventions
<Detected patterns that must be followed: naming conventions, file structure, import style, error handling approach,
state management patterns, API patterns. Be specific — "functions use camelCase, files use kebab-case, React components
use PascalCase" not "follow standard conventions".>
## Testing Patterns
<Test framework, assertion style, mocking approach (note: mock only at boundaries), file naming (_.test.ts vs
_.spec.ts), co-located vs separate, any test utilities or helpers, how to run subsets of tests.>
## Key Decisions
<Summary of relevant ADRs or architectural decisions found in docs. Link to
the actual ADR files so they can be read in full if needed.>
## Gotchas
<Non-obvious things about this project — unusual patterns, things that look wrong but are intentional, known sharp
edges, common mistakes.>
## Notes
<Anything else relevant — deployment quirks, environment setup requirements, external service dependencies, etc.>Do NOT generate a thin, generic CLAUDE.md. Every section should contain specific, actionable information derived from actually reading the codebase. If a section has nothing meaningful to say, omit it rather than filling it with generic advice.
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.

