Claude Code Config
https://github.com/jarrodwatts/claude-code-config- Category
- AI Agents
- Rank
- No. 1378Tools index
Previous survey · No. 1384 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- jarrodwatts
- GitHub
- 1.1k stars
- Date
About
Jarrod Watts' opinionated Claude Code configuration — rules, hooks, agents, skills, and commands you can fork.
What it does
This is a ready-made operating policy for Claude Code. Matching project files activate language and testing guidance, while lifecycle hooks inspect prompts and edits. Special prompt keywords inject more demanding search, analysis, or planning behavior. Global instructions also shape delegation, verification, research, and implementation decisions across sessions.
Why it's ranked here
The configuration contains useful, inspectable automation instead of prompts alone. Scoped rules reduce irrelevant guidance, hooks react at meaningful workflow points, and installation supports selective adoption. The verdict is mixed because its global behavior is strongly personal, some checks use simple heuristics, and the comment hook records part of its input in a local debug log.
What's good
The installer explicitly protects existing configuration by offering skip, overwrite, or merge choices. Rules activate only for matching TypeScript, test, or Solidity work. The comment checker covers several common languages, exempts documentation and tooling directives, and warns only when comments exceed a defined ratio. The bundled React guidance also includes parsing, validation, document generation, and evaluation-case extraction utilities.
Tradeoffs
Installing everything changes behavior across every Claude Code session, including strong assumptions about delegation and process. Common words such as search, find, analyze, and debug trigger added instructions, so ordinary prompts may receive more ceremony than intended. The comment check counts line-leading comments rather than parsing syntax, and it writes up to part of the raw hook input into a persistent local log.
How to use it well
It best suits Claude Code users who want firm defaults for TypeScript, React, testing, or Abstract chain development. Start with selective installation, then inspect global behavior and hooks before enabling them broadly. Keep the planning and research agents when their workflow matches yours. This does not replace Claude Code, a project test suite, dependency tooling, or application-specific engineering standards.
Technical notes+
README.md documents copy-paste, clone, and selective installation. INSTALL.md supplies a Claude Code prompt that fetches assets into ~/.claude, checks destination collisions, and requests skip, overwrite, or merge decisions. CLAUDE.md defines the global intent gate, task classification, delegation, and verification policy. hooks/keyword-detector.py reads JSON from stdin and emits UserPromptSubmit context for keyword-matched modes. hooks/check-comments.py handles Write, Edit, and MultiEdit events, applies a 0.25 comment ratio threshold, emits additional context, and logs diagnostic input to ~/.claude/hooks/debug.log. skills/vercel-react-best-practices/src/parser.ts parses Markdown rules; build.ts generates AGENTS.md; validate.ts checks structure and impact values; extract-tests.ts writes test-cases.json; migrate.ts splits RPG.md into rule files.
Observed
- Primary format
- Markdown configuration, supplemented by Python hooks, a shell hook, and TypeScript skill utilities.
- Install surface
- Copy-paste installation through Claude Code, full Git clone into ~/.claude, or selective component copying.
- Interface
- Claude Code global instructions, path-scoped rules, model-invoked skills, custom subagents, slash commands, and lifecycle hooks.
- Hook events
- Prompt submission, post-write or post-edit processing, and session stop are documented hook points.
- Rule scope
- Included rules target TypeScript and TSX files, TypeScript tests, general comments, and Solidity or Foundry work.
- Bundled utilities
- The React guidance skill includes Markdown parsing, validation, document generation, migration, and evaluation-case extraction scripts.
Read from README.md, hooks/check-comments.py, hooks/keyword-detector.py, skills/vercel-react-best-practices/src/build.ts, skills/vercel-react-best-practices/src/types.ts, skills/vercel-react-best-practices/src/parser.ts, skills/vercel-react-best-practices/src/migrate.ts, skills/vercel-react-best-practices/src/validate.ts, skills/vercel-react-best-practices/src/extract-tests.ts, CLAUDE.md, INSTALL.md, rules/forge.md, rules/testing.md, rules/comments.md, rules/typescript.md.
What it can do
Provide pre-configured coding rules and guidelines
Development context or coding request → Structured coding standards and best practices
Execute automated code hooks and workflows
Code changes or repository events → Automated code validation or deployment actions
Deploy specialized coding agents for different tasks
Specific development requirements → Targeted code assistance or automation
Apply pre-built coding skills and templates
Development scenarios or project types → Code scaffolding or implementation patterns
Execute custom development commands
Command parameters and project context → Code generation or project modifications
Fork and customize configuration settings
Base configuration and user preferences → Personalized Claude Code setup
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.