How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
im-hashimCLAUDE.mdim-hashim/automotive-claude-code-agents/CLAUDE.md
Automation scripts (shell) workflows/ -- End-to-end development workflows (YAML) rules/ -- Coding, safety, and security standard rules hooks/ -- Git lifecycle hooks (pre-commit, pre-push) knowledge-base/ -- Standards reference (AUTOSAR
nicepkgCLAUDE.mdnicepkg/auto-company/CLAUDE.md
性能分析+架构评审模板 | critic-munger, cto-vogels, qa-bach | ### 工程与安全 | 技能 | 能力 | 推荐角色 | |------|------|----------| | `code-review-security` | 代码审查 + 安全审计一体化 | fullstack-dhh, cto-vogels | | `security-audit` | 独立安全审计框架 | cto-vogels, devops-hightower | | `devops` | DevOps
sd0xdevCLAUDE.mdsd0xdev/sd0x-harness/CLAUDE.md
finding whose severity was mis-assigned to something that is really a security or data-integrity defect) — anything else is a `[DEVIATION]`, not a judgment call. Skill discovery: no command
AWSCLAUDE.mdawslabs/loom/CLAUDE.md
outcomes. ### Frontends - Use Typescript and shadcn, Tailwind CSS, and Vite for frontend user experience. ## Security Scans - Before committing code, verify that no credentials, tokens, or secrets are stored in files
minipuftCLAUDE.mdminipuft/claude-prompts-mcp/CLAUDE.md
three queries, `run-registry` compares it to refuse a cross-workspace handoff claim -- a security decision -- `execution-record-store` projects it, and `v_execution_status` selects it. What is true
StarRocksCLAUDE.mdStarRocks/mcp-server-starrocks/CLAUDE.md
string error messages rather than raising exceptions - Cursors are always closed in finally blocks ### Security - SQL injection prevention through parameterized queries and backtick escaping - Plotly expressions are validated using
tdiCLAUDE.mdtdi/awesome-private-ai/CLAUDE.md
Serving - Fine-Tuning & Adapters - Vector Databases & Embeddings - Agents & Orchestration - VS Code Plugins & Extensions - Privacy, Security & Governance - Models for Private Deployment - UI & Interaction Layers - Datasets & Data Prep - Learning Resources & Research
alexei-ledCLAUDE.mdalexei-led/cloud-mcp-server/CLAUDE.md
state ### Prompts (`prompts.py`) - 10+ prompt templates for common AWS tasks (Well-Architected, cost optimization, security) - Consistent pattern: Pydantic `Field(description=...)` for all parameters ## Testing Patterns - **Mocking AWS CLI:** `unittest.mock.patch
theaiautomatorsCLAUDE.mdtheaiautomators/claude-code-agentic-rag-masterclass/CLAUDE.md
calls only - Use Pydantic for structured LLM outputs - All tables need Row-Level Security - users only see their own data - Stream chat responses via SSE - Use Supabase Realtime for ingestion
cloudQuantCLAUDE.mdcloudQuant/backtrader/CLAUDE.md
line-length 100 make format-check make lint # ruff make type-check # mypy make security # bandit make quality-check # all of the above (no tests) bash scripts/optimize_code.sh # pyupgrade + isort + black
freemaCLAUDE.mdfreema/openclaw-mcp/CLAUDE.md
implements the MCP specification to expose OpenClaw capabilities as tools that Claude can invoke. ## Security Policy (CRITICAL) This is a **security-critical** MCP server. Follow these rules: ### Docker-First Deployment
icebear0828CLAUDE.mdicebear0828/clio/CLAUDE.md
diff " to read the changes. Check for: type errors, logic bugs, security issues, missing tests, style violations. If the PR is acceptable, run "gh pr review --approve
nextlevelbuilderCLAUDE.mdnextlevelbuilder/skillx/CLAUDE.md
owner, repo, skill_path?, scan? }`. GitHub ownership verified via collaborator check. Content scanned for security (risk_label: safe/caution/danger/unknown). CLI `skillx publish owner/repo [--path X] [--scan] [--dry-run]` — requires
jenticCLAUDE.mdjentic/jentic-one/CLAUDE.md
must be `type(scope): subject` with a **mandatory scope** (e.g. `feat(broker): ...`, `fix(security): ...`, `ci(ci.yml): ...`). Reserve `fix` for real bugs present on `main`; refactoring your own unshipped branch code
jtgsystemsCLAUDE.mdjtgsystems/Custom-Modes-Roo-Code/CLAUDE.md
specialized AI agent configurations for Roo Code, designed for modern software development following 2025 security-first principles and best practices. This project includes Python utilities for validation, conversion, and management
conorluddyCLAUDE.mdconorluddy/xclaude-plugin/CLAUDE.md
disclosure** - Document essentials, details on-demand - **Function size** - Prefer <50 lines, maximum 100 lines - **Security first** - Use spawn (not shell) for command execution - **Comprehensive JSDoc** - All public APIs must
MLT-OSSCLAUDE.mdMLT-OSS/FirstData/CLAUDE.md
level` is accurate and not overstated - [ ] Run `make check` to validate all checks pass ## Security Note for Contributors - Please do not paste or run commands from untrusted posts/comments. - Never include
nikolai-vysotskyiCLAUDE.mdnikolai-vysotskyi/trace-mcp/CLAUDE.md
NEVER delete code without verifying it's dead first ### Before PR / commit 1. `scan_security` { rules: ["all"] } — OWASP Top-10 vulnerability scan 2. `check_quality_gates` { scope: "changed" } — quality gate
PostHogCLAUDE.mdPostHog/duckgres/CLAUDE.md
# Claude Code Context for Duckgres This file provides context for Claude Code sessions working
CursorTouchCLAUDE.mdCursorTouch/MacOS-MCP/CLAUDE.md
style docstrings for public functions - Document Args, Returns, and Raises sections - Include type hints ## Security Considerations ⚠️ **Important**: macOS-MCP has full system access with no sandboxing. It executes real actions
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.
