claude-optimised
Guide for writing and optimizing CLAUDE.md files for maximum Claude Code performance. Use when creating new CLAUDE.md, reviewing existing ones, or when user asks about CLAUDE.md best practices. Covers structure, content, pruning, and common mistakes.
- License 许可
- MIT
- 最近更新
- 2026-06-07
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- claude-optimised
- description
- Guide for writing and optimizing CLAUDE.md files for maximum Claude Code performance. Use when creating new CLAUDE.md, reviewing existing ones, or when user asks about CLAUDE.md best practices. Covers structure, content, pruning, and common mistakes.
03 SKILL.md 正文
CLAUDE.md Optimization Guide
Write CLAUDE.md files that maximize Claude's adherence and performance.
Core Principle: Less Is More
Long CLAUDE.md = Claude ignores half of it. Critical rules get lost in noise.
For each line ask: "Would removing this cause Claude to make mistakes?"
- If no → delete it
- If Claude already does it correctly → delete it or convert to hook
What to Include
Essential (High Value)
| Section | Example |
|---|---|
| Project context | "Next.js e-commerce app with Stripe" (1 line) |
| Build/test commands | npm run test, pnpm build |
| Critical gotchas | "Never modify auth.ts directly" |
| Non-obvious conventions | "Use vi for state, not useState" |
| Domain terminology | "PO = Purchase Order, not Product Owner" |
Include Only If Non-Standard
- Branch naming (if not
feature/,fix/) - Commit format (if not conventional commits)
- File boundaries (sensitive files to avoid)
Do NOT Include
- Things Claude already knows (general coding practices)
- Obvious patterns (detectable from existing code)
- Lengthy explanations (be terse)
- Aspirational rules (only real problems you've hit)
Structure
# Project Name
One-line description.
## Commands
- Test: `npm test`
- Build: `npm run build`
- Lint: `npm run lint`
## Code Style
- [Only non-obvious conventions]
## Architecture
- [Brief, only if complex]
## IMPORTANT
- [Critical warnings - use sparingly]
Formatting Rules
- Bullet points over paragraphs
- Markdown headings to separate modules (prevents instruction bleed)
- Specific over vague: "2-space indent" not "format properly"
- IMPORTANT/YOU MUST for critical rules (use sparingly or loses effect)
File Placement
| Location | Scope |
|---|---|
~/.claude/CLAUDE.md |
All sessions (user prefs) |
./CLAUDE.md |
Project root (share via git) |
./subdir/CLAUDE.md |
Loaded when working in subdir |
.claude/rules/*.md |
Auto-loaded as project memory |
Optimization Checklist
Before finalizing:
- [ ] Under 50 lines? (ideal target)
- [ ] Every line solves a real problem you've encountered?
- [ ] No redundancy with other CLAUDE.md locations?
- [ ] No instructions Claude follows by default?
- [ ] Tested by observing if Claude's behavior changes?
Maintenance
- Run
/initas starting point, then prune aggressively - Every few weeks: "Review this CLAUDE.md and suggest removals"
- When Claude misbehaves: add specific rule
- When Claude ignores rules: file too long, prune other content
Anti-Patterns
| Don't | Why |
|---|---|
| 200+ line CLAUDE.md | Gets ignored |
| "Write clean code" | Claude knows this |
| Duplicate rules across files | Wastes tokens, conflicts |
| Theoretical concerns | Only add for real problems |
| Long prose explanations | Use bullet points |
Example: Minimal Effective CLAUDE.md
# MyApp
React Native app with Expo. Backend is Supabase.
## Commands
- `pnpm test` - run tests
- `pnpm ios` - run iOS simulator
## Style
- Prefer Zustand over Context
- Use `clsx` for conditional classes
## IMPORTANT
- NEVER commit .env files
- Auth logic lives in src/lib/auth.ts only
~15 lines. Covers what Claude can't infer. Nothing more.
04 同仓库更多 Skills
crowd-prompting
A marketplace where AI agents improve prompts, system instructions, tool descriptions, and other text-based content with domain expertise from real-world operations — and earn tokens for valuable contributions.
ceaser
Interact with the Ceaser privacy protocol on Base L2 using the ceaser-mcp MCP tools. This skill uses the ceaser-mcp npm package for ALL operations -- shield, unshield, note management, and protocol queries. All ceaser tool calls use CLI subcommands (npx -y ceaser-mcp <subcommand>). Ceaser is a privacy-preserving ETH wrapper using Noir/UltraHonk zero-knowledge proofs on Base L2 (chain ID 8453).
ceaser-send
Fully automated private ETH transfer via Ceaser Protocol on Base L2 using the ceaser-mcp MCP tools. This skill uses the ceaser-mcp npm package for ALL operations -- all ceaser tool calls use CLI subcommands (npx -y ceaser-mcp <subcommand>). WARNING -- creates observable on-chain links between user wallet and hot wallet (see Privacy Warning). Generates an ephemeral hot wallet with BIP-39 mnemonic recovery, user funds it, agent signs and broadcasts the Shield TX automatically. Exactly one manual step required (funding the hot wallet with ETH). Uses Noir/UltraHonk zero-knowledge proofs.
dilemm-ai
Compete in DilemmAI, the prisoner's dilemma AI arena at dilemm.ai. Use when an OpenClaw agent wants to create an account, design and submit strategy prompts for their bot, enter matchmaking, analyze historical matches and leaderboards, and iterate on bot strategy. Requires headless browser automation (no API available yet).
moldium
Post and manage content on the Moldium blog platform. Triggered by "post to Moldium", "write a blog post", "publish an article", etc.
arxiv-paper-reviews
Interact with arXiv Crawler API to fetch papers, read reviews, submit comments, search papers, and import papers. Use when working with arXiv papers, fetching paper lists by date/category/interest, viewing paper details with comments, submitting paper reviews, searching papers by title, or importing papers from arXiv URLs via API at http://weakaccept.top:8000/.