How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
ruvnetCLAUDE.mdruvnet/agentic-flow/CLAUDE.md
performance-engineer` ### Swarm Coordination `hierarchical-coordinator`, `mesh-coordinator`, `adaptive-coordinator` ### GitHub & Repository `pr-manager`, `code-review-swarm`, `issue-tracker`, `release-manager` ### SPARC Methodology `sparc-coord`, `sparc-coder`, `specification`, `pseudocode`, `architecture
liangdabiaoCLAUDE.mdliangdabiao/amazon-sorftime-research-MCP-skill/CLAUDE.md
CLAUDE.md This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository. ## Project Overview Amazon competitive intelligence and category selection analysis workspace using Sorftime
gotempshCLAUDE.mdgotempsh/temps/CLAUDE.md
other detail that identifies whose system produced a payload must never appear in code, comments, test fixtures, commit messages, branch names, file names, PR titles, PR descriptions, PR comments
23blocks-OSCLAUDE.md23blocks-OS/ai-maestro/CLAUDE.md
must be propagated: Browser → WebSocket → PTY → tmux - Input/output is binary-safe (supports ANSI escape codes, Unicode, etc.) When working with terminal components: - xterm.js handles rendering only - it doesn't know
StripeCLAUDE.mdstripe/link-cli/CLAUDE.md
CLAUDE.md This file provides guidance to Claude Code when working with code in this repository. ## Project Overview Link CLI — lets agents get secure, one-time-use payment credentials from
MongoDBCLAUDE.mdmongodb/docs/CLAUDE.md
task will clearly exceed these sizes, propose splitting it into multiple independently reviewable PRs *before* writing code, and call out the split in your plan. - If a size warning fires
MongoDBCLAUDE.mdmongodb/docs/platform/CLAUDE.md
code already says. Keep it short and specific. If a comment doesn't clearly earn its place, omit it. The goal is signal-dense PRs that are easy to review
alirezarezvaniCLAUDE.mdalirezarezvani/claude-code-tresor/CLAUDE.md
November 19, 2025 (v2.7.0) ## 🏗️ Architecture ``` claude-code-tresor/ ├── skills/ # 8 Autonomous Skills (v2.0+) │ ├── development/ # code-reviewer, test-generator, git-commit-helper │ ├── security/ # security-auditor, secret-scanner, dependency-auditor │ └── documentation
Kc1tCLAUDE.mdKc1t/alethe-agents/CLAUDE.md
Windows-first** desktop app that organizes, operates, and resumes multiple coding agents (Claude Code, Codex, OpenCode) and shells in parallel, inside a persistent workspace with real terminals (PTYs), layouts, themes
ethanplusaiCLAUDE.mdethanplusai/jarvis/CLAUDE.md
warns, and then needs `JARVIS_ALLOWED_ORIGINS`. ## The Run Pipeline Nothing spawns Claude Code outside `RunExecutor`. Two invariants govern it: 1. **A run always reaches a terminal state.** `succeeded`, `failed
ethanplusaiCLAUDE.mdethanplusai/jarvis/jarvis_home/CLAUDE.md
spec written to disk and approved by voice, then a real Claude Code session driven through plan, review and execute. - You watch every Claude Code session on the machine
CryptoDmitryCLAUDE.mdCryptoDmitry/Ghost-Agent/CLAUDE.md
kcov (informational only) - Uploads coverage artifacts 2. **claude.yml** - Claude Code GitHub Actions integration - Automated code review capabilities 3. **claude-code-review.yml** - PR code review workflow - Automated review on pull requests ### Coverage Note
jwangkunCLAUDE.mdjwangkun/claude-for-financial-services-cn/CLAUDE.md
wind__*`, and `mcp__china-news__*` syntax. - Stock codes follow A-share conventions: 6-digit codes, no exchange prefix (e.g., "600519", "000001"). ## Agents (4 China-market) - `china-pitch-agent` — A-share
MicrosoftCLAUDE.mdmicrosoft/TypeAgent/ts/CLAUDE.md
stored in `~/.typeagent/` - Tracing via the `debug` package — enable with `DEBUG=typeagent:*` env var ## Code Review Guidelines Use these criteria both when **writing** changes and when **reviewing** code (your
kgoedeckeCLAUDE.mdkgoedecke/doop/CLAUDE.md
which `.context/INDEX.md` links. Always use the `clean-code:typescript` skill when writing or reviewing TypeScript code in this repo. ## First-time setup after clone ```bash bun install
alinaqiCLAUDE.mdalinaqi/maggy/CLAUDE.md
Grok 4.3 (xAI) | $5 / $15 | Competitor intel, CKG, deep reasoning | | 10 | Codex | varies | Code review, bulk generation | | 11 | Claude Sonnet/Opus | $3-5 / $15-25 | Quality-critical, security, architecture | ### Use qwen3
alinaqiCLAUDE.mdalinaqi/maggy/templates/CLAUDE.md
line) - Use self-deprecating humor about AI when it fits ("I've reviewed 500 lines of code and my only complaint is that I can't drink coffee while doing
trailofbitsCLAUDE.mdtrailofbits/coop/CLAUDE.md
Before opening, run the [`closeout-review`](.agents/skills/closeout-review/SKILL.md) skill on the working diff (a `PreToolUse` hook gates `gh pr create` on it). Describe what the code does now — plain, factual language
gmickelCLAUDE.mdgmickel/flow-next/CLAUDE.md
command discovery. Invoke the selected skill with `flow-next:flow-next- ` on Claude Code; follow its task-specific contract. This is a repo-customized block. Keep maintainer additions outside
rshankrasCLAUDE.mdrshankras/claude-code-apple-skills/CLAUDE.md
auditor`) skills | ### Generator vs Advisory vs Workflow Skills - **Advisory skills** (ios/, macos/, release-review/): Review code and provide recommendations - **Generator skills** (generators/): Produce production-ready Swift code with templates - **Workflow
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.
