How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
ElasticCLAUDE.mdelastic/integrations/CLAUDE.md
Code guidance for elastic/integrations This repo has companion Claude Code skills in [`elastic/integration-skills`](https://github.com/elastic/integration-skills) that cover building, reviewing, and maintaining Elastic integration packages. If the skills aren
agents-universeCLAUDE.mdagents-universe/agents-universe/CLAUDE.md
owns the turn). 8. **Skill types** — `guidance` (LLM instructions), `template` (code templates), `executable` (runnable code blocks), `composite` (chains other skills). 9. **Workflow definitions** — Same format as skills; files
DatabricksCLAUDE.mddatabricks/databricks-agent-skills/CLAUDE.md
Databricks Agent Skills Skills for AI coding assistants (Claude Code, etc.) that provide Databricks-specific guidance. ## Structure ``` skills/ ├── databricks-core/ # core skill: CLI, auth, data exploration │ ├── SKILL.md │ └── *.md (references) └── databricks
K-Dense-AICLAUDE.mdK-Dense-AI/science-superpowers/CLAUDE.md
work, designing analyses, pre-registering predictions, running reproducible analyses, investigating anomalies, verifying results, and reviewing them adversarially. The architecture is reused from [Superpowers](https://github.com/obra/superpowers): a session-start bootstrap
brycewang-stanfordCLAUDE.mdbrycewang-stanford/StatsPAI/CLAUDE.md
CLAUDE.md — StatsPAI > Claude Code / AI agent 在本仓库工作的指引。动手前请通读。 --- ## 1. 项目定位 **StatsPAI** 的目标是**面向 Agent 设计，适合人类以及 agent 进行调用，并致力于超越 Stata、R、以及老的 Python 生态，成为全世界最好的因果推断与实证分析工具**。 做法有三条： 1. **一次 `import statspai
AnthropicCLAUDE.mdanthropics/anthropic-sdk-csharp/CLAUDE.md
release flow breaks the sync. Rebase onto the current default branch before asking for review — a stale base shows up as dozens of unrelated "changed" files. - Package versions (`VersionPrefix`, `.release-please-manifest.json
textcortexCLAUDE.mdtextcortex/claude-code-sandbox/CLAUDE.md
integration. The tool creates safe sandboxed environments where Claude can execute commands and make code changes without affecting the host system. ## Common Development Commands ### Build and Development - `npm run build
Encod3d-SecCLAUDE.mdEncod3d-Sec/TORCH/CLAUDE.md
driver scripts in `scripts/burp/`, host setup in `setup/burp/`), and the standalone `wiki/`, `research/`, `disclosure/`, `code-review/`. They load on demand via the Skill tool (descriptions in the `/skills` picker), discovered
niaka3dayoCLAUDE.mdniaka3dayo/agent-skills-vrc-udon/CLAUDE.md
skills/*/SKILL.md` | Skill definitions with YAML frontmatter | | `skills/*/rules/*.md` | Constraint rules for AI code generation | | `templates/*.md` | AI tool config files distributed to end users | ## Testing ```bash # Verify
sepivipCLAUDE.mdsepivip/SeekerClaw/CLAUDE.md
networks) - Disabled skills: browser, canvas, nodes, screen ## Memory Preservation (CRITICAL) > **RULE: App updates and code changes MUST NEVER affect user memory.** The agent's memory is sacred. These files live
gendigitalincCLAUDE.mdgendigitalinc/sage/CLAUDE.md
calling the evaluator. ## Pre-PR Checklist If you use Claude Code, run `/simplify` before creating a pull request to review changed files for clarity, consistency, and maintainability improvements. Apply
clay-goodCLAUDE.mdclay-good/OpenLore/CLAUDE.md
tool | | "What's unreachable / what dies if I delete X?" | `find_dead_code` — cross-language reachability (candidates) | | Reviewing a change: structural delta + stale callers | `structural_diff` | | "Did my diff escape
ucsandmanCLAUDE.mducsandman/DashClaw/CLAUDE.md
capability to the Claude Code plugin, mirror it for Codex and **Hermes** (`plugins/`); those parity gaps have been flagged. - **On PR/spec reviews, trust only what you read** — don't pass
ZSeven-WCLAUDE.mdZSeven-W/dsh-ios/demo-video/CLAUDE.md
captions/subtitles → `/embedded-captions`.) - `/pr-to-video` — a GitHub PR (URL / `owner/repo#N` / "this PR") → 30-90s code-change explainer (changelog / feature reveal / fix / refactor). - `/motion-graphics` — a short (typically under 10s) design
skilld-devCLAUDE.mdskilld-dev/skilld/CLAUDE.md
same Skill files directly without the Harness. Direct runs remain user reviewed. Harness and CI runs enforce strict output checks. ## Architecture - `crates/skilld-core`: shared Rust domain types and Agent target rules
theNetworkChuckCLAUDE.mdtheNetworkChuck/claude-phone/CLAUDE.md
name]` | Build with TDD | | `/feature ship` | Review and merge | | `/test` | Run tests | | `/fix [N]` | Fix GitHub issue #N | | `/investigate [problem]` | Debug without changing code | ## API Endpoints ### Voice App (port
ai-analyst-labCLAUDE.mdai-analyst-lab/ai-analyst/CLAUDE.md
validates dependencies and computes execution tiers. The version-3 controller runs one fresh Claude Code process per worker, sequentially, with explicit inputs and per-attempt outputs. It stops on unavailable
AnttiHeroCLAUDE.mdAnttiHero/lavern/CLAUDE.md
engagement` — Intake and team selection - `verification` — Standalone document verification pipeline - `tabulate` — Tabular multi-document review (one row per doc, every cell cited) - `src/workflows/executor.ts` — Generic workflow runner with soul + personality injection
Prorise-coolCLAUDE.mdProrise-cool/Claude-Code-Multi-Agent/CLAUDE.md
MODE: PLAN] (计划与规范) **视角**: 📋 技术主管 (Tech Lead) + 测试工程师 (TE) * **目标**: 生成原子级的可执行计划。 * **禁止**: 编写大段示例代码 (Example Code)、模棱两可的描述。 * **强制动作序列**: 1. **可行性预检**: 用 `Exa AI` 查阅库版本兼容性。 2. **任务拆解**: 生成 **带编号的原子检查清单 (Checklist)**，精确到文件名
theexperiencecompanyCLAUDE.mdtheexperiencecompany/gaia/CLAUDE.md
same human, and cross-surface funnels silently stop joining — the failure is invisible in code review and only shows up as wrong numbers. - Authenticated API requests get this for free
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.
