mlx-serve / app
ddalcu/mlx-serve/app/CLAUDE.md
# Swift macOS app (MLX Core) — AI context Compressed rules. Full war stories: `../docs/gotchas/app.md`. Full
How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
ddalcu/mlx-serve/app/CLAUDE.md
# Swift macOS app (MLX Core) — AI context Compressed rules. Full war stories: `../docs/gotchas/app.md`. Full
stripe/stripe-cli/CLAUDE.md
rather than `/tmp` — `bin/` is gitignored and within the project's trusted directory. macOS security may block execution of binaries built outside the project tree. ## Project Structure - `cmd/stripe/main.go` - Entry point
Steady11 mentions in 2 weeks
No other project uses it yet
joshuayoes/ios-simulator-mcp/CLAUDE.md
xcrun simctl` and Facebook's `idb` commands - Validates all inputs with Zod schemas - Implements security best practices with `--` argument separation - Supports environment-based tool filtering via `IOS_SIMULATOR_MCP_FILTERED
Lumiwealth/lumibot/CLAUDE.md
behavior description ### Fixed - Bug fix description ### Deprecated - Feature being phased out ### Removed - Removed feature ### Security - Security-related changes ``` ### Pre-Deployment Checklist Before any deployment: - [ ] Reviewed git commits since last version
photon-hq/spectrum-ts/.claude/CLAUDE.md
Prefer simple conditionals over nested ternary operators - Group related code together and separate concerns ### Security - Add `rel="noopener"` when using `target="_blank"` on links - Avoid `dangerouslySetInnerHTML` unless absolutely necessary
yakhyo/uniface/CLAUDE.md
full suite before submitting changes. ## Pre-commit Pre-commit hooks handle formatting, linting, security checks, and notebook output stripping. Always run: ```bash pre-commit install pre-commit run --all-files
anthropics/anthropic-sdk-typescript/CLAUDE.md
middleware code, where a failure tears down the caller's response. - A permission or security decision is the exception. It treats an unrecognized value as a denial. - When API output
opensearch-project/OpenSearch-Dashboards/CLAUDE.md
with autofix node scripts/eslint --fix # Type check yarn typecheck # Run Cypress functional tests (no security plugin) yarn cypress:run-without-security # Run a specific Cypress test yarn cypress:run-without
WuKongOpenSource/AI_CRM/CLAUDE.md
enabled` 默认 `false`。同一用户重新登录会写 kickout 标记并删除旧令牌,旧令牌的后续请求随即失效;置 `true` 放开多端并发登录。 - **每请求生命周期**:校验令牌 → 写入 Spring Security 上下文与 `TenantContextHolder` → 在 `finally` 中**同时**清理 `TenantContextHolder` 与 `DataPermissionHolder`(线程池复用时防 ThreadLocal 泄漏)。无请求上下文的代码(注册
ipxe/ipxe/CLAUDE.md
AGENTS.md`](AGENTS.md) — follow them. This file adds only the Claude Code-specific pieces. - **Security review:** for vulnerability-hunting work, invoke the **ipxe-security-review** skill (`.claude/skills/ipxe-security-review/`). It encodes the sweep
No other project uses it yet
trycompai/comp/CLAUDE.md
Monorepo Structure ``` apps/ api/ # NestJS API (auth, RBAC, business logic) app/ # Next.js frontend (compliance + security products) portal/ # Employee portal packages/ auth/ # RBAC definitions (permissions.ts) — single source of truth db/ # Prisma
No other project uses it yet
doobidoo/mcp-memory-service/CLAUDE.md
README. ## Development Guidelines ### Quality gates Three layers: pre-commit (under 5s, Groq/Gemini complexity plus security, blocks on complexity above 8 or a security finding), the PR gate `bash scripts/pr/pre_pr_check.sh
Steady5 mentions in 2 weeks
No other project uses it yet
YosemiteCrew/Yosemite-Crew/CLAUDE.md
never skip Identify and fix Sonar findings **locally before pushing** — never let issues or security hotspots first surface on the PR. - Before pushing ANY change, run the local Sonar workflow
ParabolInc/parabol/.claude/CLAUDE.md
checks live in permissions.ts (graphql-shield), not inline in resolvers. Never trust client-supplied security facts (granted scopes, team ids) — derive them server-side. Every new GraphQL field needs explicit
Steady7 mentions in 2 weeks
No other project uses it yet
microsoft/PowerApps-Samples/CLAUDE.md
ImageOperations**: Image column handling - **Schema**: Metadata operations, entity creation - **Search**: Dataverse search, elastic search - **Security**: Role management, sharing, field security - **ServiceClient**: Modern client usage patterns - **Solutions**: Solution management and deployment
ranuts/document/CLAUDE.md
次骰子,而骰子是歪的:runner 首选的 `azure.archive.ubuntu.com` 返回 `Ign:`,退回公网 archive 后整个停摆(`noble-security InRelease` 之后零字节)。 本周四个 job 因此被 6 小时上限杀掉,加固之后又有两次各烧 17 分钟。 现在缓存命中直接 `exit 0`(连 apt 配置都不写);`PLAYWRIGHT_INSTALL
beikeshop/beikeshop/CLAUDE.md
level `lang/` 4. **Cache**: Clear cache after config changes with `php artisan config:clear` ## Security - Never commit `.env` file - Change default admin password in production - Use HTTPS in production environment
WorldFlowAI/everything-claude-code/examples/CLAUDE.md
Unit tests for utilities - Integration tests for APIs - E2E tests for critical flows ### 4. Security - No hardcoded secrets - Environment variables for sensitive data - Validate all user inputs - Parameterized queries only
23 projects use a version of this
vstorm-co/full-stack-ai-agent-template/CLAUDE.md
backend/ │ ├── app/ │ │ ├── main.py # FastAPI app with lifespan │ │ ├── api/ # Routes, deps, exception handlers │ │ ├── core/ # Config, security, middleware │ │ ├── db/ # Models, session management │ │ ├── schemas/ # Pydantic request/response models │ │ ├── repositories/ # Data access layer │ │ ├── services/ # Business
bcgit/bc-csharp/CLAUDE.md
source tree; the FIPS distribution is separate. Root namespace: `Org.BouncyCastle`. NuGet package id: `BouncyCastle.Cryptography`. Security issues: see [SECURITY.md](SECURITY.md) — report privately to feedback-crypto@bouncycastle.org rather than filing a public issue. ## Solution
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.