How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
cyanheadsCLAUDE.mdcyanheads/mcp-ts-core/CLAUDE.md
CUSTOM_KV']], // KV/R2/D1 → globalThis onScheduled: async (controller, env, ctx) => { /* cron */ }, }); ``` Per-request `McpServer` factory (security: SDK GHSA-345p-7cg4-v4c7). Requires `compatibility_flags = ["nodejs_compat"]` and `compatibility_date
cyanheadsCLAUDE.mdcyanheads/mcp-ts-core/templates/CLAUDE.md
report of issues and pain points 7. **Run `devcheck`** — lint, format, typecheck, and security audit 8. **Run the `security-pass` skill** — audit handlers for MCP-specific security gaps: output injection
nwiizoCLAUDE.mdnwiizo/ccswarm/CLAUDE.md
Builtin facets Personas: planner, coder, reviewer, researcher, supervisor, ai-antipattern-reviewer. Policies: coding, review, security, testing. Knowledge: (user-provided under `.ccswarm/facets/knowledge/*.yaml`). ## Pipeline learnings - タスク記述は簡潔に(500語以下)。長いとimplementがタイムアウト。 - `{task}`, `{plan_output
sorkilaCLAUDE.mdsorkila/lockpaw/CLAUDE.md
creates an entry — **grant the host terminal (e.g. Ghostty) → Finder under System Settings → Privacy & Security → Automation and re-run**; v1.1.1's branded DMG was built from Claude Code this
cloudnative-coCLAUDE.mdcloudnative-co/claude-code-starter-kit/CLAUDE.md
summary report. The comparison basis is always "real `~/.claude` vs sim dir result". ## Security Hardening `config/permissions.json` keeps high-risk tools out of `permissions.allow` so code execution, network fetches, environment reads
Comfy-OrgCLAUDE.mdComfy-Org/comfy-claude-prompt-library/CLAUDE.md
Perfect documentation 2. Simple solution → Complex solution (only if needed) 3. Evidence/Data → Opinion/Preference 4. Security first → Performance → Features 5. Explicit > Implicit ## Ambiguity Resolution When requests are vague ("fix it", "make
ever-worksCLAUDE.mdever-works/ever-works/CLAUDE.md
users need to set up login, sign-in, sign-up, credential authentication, or password security with Better Auth. - `.claude/skills/email-and-password-best-practices/SKILL.md` ## Design Thinking Create distinctive, production-grade frontend interfaces with high design
AnastasiyaWCLAUDE.mdAnastasiyaW/codex-claude-code-config/CLAUDE.md
what we are doing, (2) only relevant state, (3) constraints - Example: "Do a security review. Context: files X, Y, Z. Check: OWASP top 10. Format: JSON with severity." ## Structured Reasoning
cyanheadsCLAUDE.mdcyanheads/pubmed-mcp-server/CLAUDE.md
report of issues and pain points 7. **Run `devcheck`** — lint, format, typecheck, and security audit 8. **Run the `security-pass` skill** — audit handlers for MCP-specific security gaps: output injection
MohamedAbdallah-14CLAUDE.mdMohamedAbdallah-14/unslop/CLAUDE.md
attention anchor. ### `hooks/unslop-statusline.sh` — Statusline badge Reads flag file at `$CLAUDE_CONFIG_DIR/.unslop-active`. Security hardened: refuses symlinks, caps read at 64 bytes, strips non-alphanumeric characters, whitelists valid modes. Outputs colored
NachoSEOCLAUDE.mdNachoSEO/claudegram/CLAUDE.md
src/utils/` (download, sanitize, file-type) - Validate external input (URLs, file content) using existing helpers ## Security Checklist Before committing changes that handle external input: - [ ] URL protocol validation using `isValidProtocol()` from `src/utils/download.ts
bobmatnycCLAUDE.mdbobmatnyc/claude-mpm/CLAUDE.md
duetto) → `make release-patch|minor|major` → `make release-publish`. Never call `./scripts/publish_to_pypi.sh` directly. --- ## 🔴 Security: credential files Credential/identity files (e.g. `scripts/lib/gh_identity.sh`) must **never** be pushed directly to `main` — always open
CranotCLAUDE.mdCranot/agi-in-md/CLAUDE.md
Sonnet vanilla (7/10) by 2+ points. - **Claim prism is domain-sensitive.** 9.5 on AuthMiddleware (security-adjacent code has clearer assumptions to invert; found genuine multi-org permission escalation vulnerability
forefyCLAUDE.mdforefy/.context/CLAUDE.md
This repo contains instruction files for using and conducting security reviews, do not confuse this `CLAUDE.md` from other copilot-instructions files existing in this workspace. For every change
nkmr-jpCLAUDE.mdnkmr-jp/prompt-line/CLAUDE.md
src/renderer/`): UI and user interactions with 13+ specialized managers. See `src/renderer/CLAUDE.md` - **Preload Script** (`src/preload/preload.ts`): Secure context bridge with whitelisted IPC channels - **IPC Handlers** (`src/handlers/`): 9 specialized files, 52 IPC channels
nkmr-jpCLAUDE.mdnkmr-jp/prompt-line/native/CLAUDE.md
Native Tools Module 4 native macOS Swift tools replacing AppleScript for better performance, security, and reliability. ## Tools | Tool | Purpose | CLI Example | |------|---------|-------------| | `window-detector` | Window bounds & app detection | `window-detector current
pgplexCLAUDE.mdpgplex/pgconsole/.claude/claude.md
Full table metadata (size, row count, etc.) - `GetIndexes`, `GetConstraints`, `GetTriggers` - Table structure - `GetPolicies`, `GetGrants` - Security info - `GetMaterializedViews`, `GetFunctions`, `GetProcedures` - Object lists - `GetFunctionInfo` - Function/procedure details with definition **AIService** (`/server/services/ai-service.ts`) - `ListAIProviders` - List
DjangoPengCLAUDE.mdDjangoPeng/agentic-ai/CLAUDE.md
README. Main areas: - `openclaw-infra/`, `openclaw-im/`, `openclaw-models/`: deployment, IM integration, model/provider, security, and troubleshooting guides for OpenClaw. - `openclaw-soul/`, `openclaw-heartbeat/`, `openclaw-skills/`: templates and examples for SOUL/AGENTS
DjangoPengCLAUDE.mdDjangoPeng/agentic-ai/security-guardian/CLAUDE.md
guidance to Claude Code (claude.ai/code) when working with code in this repository. > 本文件只管 `security-guardian/` 子项目。仓库根 `agentic-ai/CLAUDE.md` 是课程总览，与本文件并存。 ## 这个项目是什么 第 20 节《企业级数字员工的安全审计与生产治理》配套项目：一个面向
maxvaegaCLAUDE.mdmaxvaega/skillkit/CLAUDE.md
pattern (metadata loading → on-demand content) - Framework integrations (LangChain, LlamaIndex, CrewAI, Haystack, Google ADK) - Security features (path traversal prevention, permission checks) - Model-agnostic design supporting Claude, GPT, Gemini, and open
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.
