How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
BelkinsCLAUDE.mdBelkins/ai-dive-deep/vault-starter/CLAUDE.md
CLAUDE.md Working memory for every AI session in this vault. Loaded into Claude Code / Cowork on wake-up. Keep this under 100 lines. Every line costs tokens on every turn
databrickslabsCLAUDE.mddatabrickslabs/ontobricks/CLAUDE.md
switches) on top of those rules. ### Skills (auto-invoked by description) | Skill | Trigger phrases | |-------|-----------------| | `code-review` | "code review", "review the code", reviewing a feature/PR/branch | | `refactoring` | "refactor", restructure, clean up, simplify
receptronCLAUDE.mdreceptron/mulmoclaude/CLAUDE.md
destination, not in `os.tmpdir()` - Package exports: include `"require"` and `"default"` conditions (Docker CJS mode) ### Code style - Functions under 20 lines; split into smaller functions if needed - `const` over `let`; never
coleam00CLAUDE.mdcoleam00/remote-agentic-coding-system/.agents/examples/claude-telegram-bot/CLAUDE.md
Research** → Use knowledge base (see RAG workflow below) 4. **Implement** → Write code based on research 5. **Review** → `manage_task("update", task_id="...", status="review")` 6. **Next Task** → `find_tasks(filter
coleam00CLAUDE.mdcoleam00/remote-agentic-coding-system/CLAUDE.md
review` on PR #15 2. Fetch PR diff: `gh pr diff 15` 3. AI: Review code, generate feedback 4. `gh pr review 15 --comment -b "feedback"` **Remote Development (Telegram/Slack
LangChainCLAUDE.mdlangchain-ai/langchain-aws/CLAUDE.md
test file uv run --group test pytest tests/unit_tests/test_specific.py ``` ```bash # Lint code make lint # Format code make format # Type checking uv run --group lint mypy . ``` #### Key config files - pyproject.toml: Main workspace
tingly-devCLAUDE.mdtingly-dev/tingly-box/CLAUDE.md
review. New feature design and PR review should be explicitly checked against these principles; deviations require a clear justification. ## Frontend API related code is generated by swagger. For each
kousenCLAUDE.mdkousen/claude-code-training/CLAUDE.md
recommended bridge pattern when a project needs to expose the same guidance to both Claude Code and AGENTS.md-aware tools without duplicating content. ## Project Structure & Standards ### Core Course Materials - **`slides.md`**: Main
automagik-devCLAUDE.mdautomagik-dev/genie/CLAUDE.md
file) bun run typecheck # tsc --noEmit bun run lint # biome check . bun run dead-code # bunx knip (has pre-existing false positives for biome/commitlint/husky) bun test # All tests bun test
automagik-devCLAUDE.mdautomagik-dev/genie/skills/refine/prompts/claude.md
keep the lead working. - **Coding:** favor targeted edits, requested scope, and tests sized to behavior and repository conventions. Frame a general correctness review as looking for bugs; preserve explicit compilation
AnthropicCLAUDE.mdanthropics/code-migration-kit-with-claude-code/CLAUDE.md
manual for Claude Code sessions in this repository You are working inside the Claude Code Migration Kit, or inside a repository that has adopted it. Read this before acting
Charlie85270CLAUDE.mdCharlie85270/Dorothy/CLAUDE.md
native memory** (`~/.claude/projects/*/memory/`) via the Memory page. No custom storage — reads real Claude Code memory files. Project dir names use path-as-folder-name encoding (slashes → dashes). ## Memory
raysonmengCLAUDE.mdraysonmeng/agent-bridge/CLAUDE.md
than doing everything yourself. ### Capability comparison | Capability | Claude (you) | Codex | |---|---|---| | Architecture & planning | Strong | Moderate | | Code review & analysis | Strong | Strong | | Sandboxed code execution | No | Yes | | File editing & refactoring | Yes (via tools
zippoxerCLAUDE.mdzippoxer/subtask/CLAUDE.md
Lead vs Worker **Lead** (e.g., Claude Code, Codex, OpenCode): - Runs `subtask` CLI commands - Sees CLI output (`list`, `show`, errors) - Drafts tasks, sends prompts, reviews work - Reads task folders and history.jsonl
Fergana-LabsCLAUDE.mdFergana-Labs/stash/CLAUDE.md
comment is to explain information that is not obvious from reading the code. A good rule of thumb: if your comment is an action (eg: sets the diff viewer open
NVIDIACLAUDE.mdNVIDIA/nvidia-resiliency-ext/CLAUDE.md
Workflow - All changes must begin with a tracked issue, approved by NVRx engineers before code review starts. - Contributions go through personal forks; do not push directly to upstream. - Prefix
TheDecipheristCLAUDE.mdTheDecipherist/claude-code-mastery-project-starter-kit/CLAUDE.md
single-line changes, obvious bugs - One Claude writes the plan. You review it as the engineer. THEN code. ### Step Naming — MANDATORY Every step in a plan MUST have a consistent
zhukunpenglinyutongCLAUDE.mdzhukunpenglinyutong/ai-max/examples/CLAUDE.md
可选 DEBUG=false ``` ## 可用命令 - `/aimax:tdd` - 测试驱动开发工作流 - `/aimax:plan` - 创建实现计划 - `/aimax:code-review` - 审查代码质量 - `/aimax:build-fix` - 修复构建错误 ## Git 工作流 - 约定式提交：`feat:`、`fix:`、`refactor:`、`docs:`、`test:` - 永不直接提交到 main 分支
Get-Concord-AICLAUDE.mdGet-Concord-AI/concord-mcp/CLAUDE.md
concord` CLI that gives coding agents shared work-state: agents **register their presence**, **claim work**, share typed **task context**, leave **handoffs**, and generate **review packets** before opening PRs. SQLite
affaan-mCLAUDE.mdaffaan-m/claude-swarm/.claude/CLAUDE.md
Claude Swarm Multi-agent orchestration for Claude Code built on the Agent SDK. ## Architecture ``` src/claude_swarm/ cli.py - Click CLI entry point (main group + sessions/replay subcommands) types.py - Core dataclasses: SwarmTask, SwarmPlan, SwarmAgent
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.
