im-not-ai
epoko77-ai/im-not-ai/CLAUDE.md
report`면 사람 검토 권고. 수렴 판정은 LLM 재탐지가 아니라 `verify_gates.py`(결정적, exit code)가 내린다. `verify_change_rate.py`는 기존 연동을 위해 보존한다. ## 에이전트 구성 (9개 — 역할 구분 필수) **런타임 3종
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.
epoko77-ai/im-not-ai/CLAUDE.md
report`면 사람 검토 권고. 수렴 판정은 LLM 재탐지가 아니라 `verify_gates.py`(결정적, exit code)가 내린다. `verify_change_rate.py`는 기존 연동을 위해 보존한다. ## 에이전트 구성 (9개 — 역할 구분 필수) **런타임 3종
srizzon/git-city/CLAUDE.md
ONLY change grants/policies — never touch the data itself - Always propose SQL as code for the user to review, never execute destructive SQL directly ## Environments - Three environments, one-directional flow: local
aiogram/aiogram/CLAUDE.md
browsing. - Read graphify-out/GRAPH_REPORT.md only for broad architecture review or when query/path/explain do not surface enough context. - After modifying code, run `graphify update .` to keep the graph current
No other project uses it yet
areal-project/AReaL/CLAUDE.md
Trigger | | --------------------------- | ----------------------------------------- | ------------------------------------------------------------------- | | `planner` | Implementation planning | Before multi-file changes, new features, or architectural decisions | | `simple-code-reviewer` | Quick code quality checks | After code changes, before committing | | `code-verifier` | Formatting/linting/tests | After code
Galaxy-Dawn/claude-scholar/CLAUDE.md
writing -> `review-response` - Bound research repo knowledge maintenance -> `obsidian-project-kb-core` For coding, debugging, architecture, review, and verification tasks, prefer the matching development skill instead of improvising. --- ## Bound Repo
pytorch/torchtitan/.claude/CLAUDE.md
After model changes, ensure HF checkpoints still load correctly. Document reasons for model changes. ## Code Review Use the `pr-review` skill to review PRs (`/pr-review <PR#>` or `/pr-review branch`); `@claude
shader-slang/slang/CLAUDE.md
this layer owns the logic, with a code trace from producer to consumer. ### Code Style and Review Conventions Recurring review feedback distilled into rules — following them avoids review round-trips
shader-slang/slang/extras/claude.md
Amend A fresh agent reviews the commit as if it were a real code reviewer: checks correctness, style, edge cases, test coverage, and regressions. It also reviews all documentation changes
coleam00/ottomator-agents/claude-agent-sdk-demos/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
zenml-io/zenml/CLAUDE.md
Before opening a PR or making a large commit**, always run `/simplify` to review changed code for reuse opportunities, quality issues, and efficiency improvements. Fix any issues it finds before
phan/phan/CLAUDE.md
file paths in `.expected` files - **Line numbers**: Critical for matching expected warnings to actual code lines - **Issue types**: Must match exactly (e.g., `PhanTypeInvalidCallableMethodName`) - **Error messages**: Complete message text must match
Steady4 mentions in 2 weeks
No other project uses it yet
weave-os/router/CLAUDE.md
write. - **Non-tautological tests** — every test must assert behavior that breaks if prod code deleted. ## Layer model and import rules Three concentric layers. Imports flow inward only. ``` +-------------------------------------------------------------------+ | cmd/router/main.go (composition root
Surging3 mentions in 2 weeks
1 project uses a version of this
unkeyed/unkey/CLAUDE.md
reasoning from the code alone. - When debugging a failure, start from the incident or issue Polylane recorded: it carries the evidence an investigation already gathered. - Polylane reviews pull requests
cloudflare/vibesdk/CLAUDE.md
this repository. ## Communication Style - Be professional, concise, and direct - Do NOT use emojis in code reviews, changelogs, or any generated content. You may use professional visual indicators or favor markdown
breaking-brake/cc-wf-studio/CLAUDE.md
becomes the CHANGELOG entry) ``` This writes a `.changeset/ .md` file — commit it alongside your code. **You never hand-edit `CHANGELOG.md`; Changesets generates it.** If a PR genuinely needs no release
streamyfin/streamyfin/CLAUDE.md
writing code in that area. | Document | Read it when | | --- | --- | | [docs/conventions/constants.md](docs/conventions/constants.md) | Adding a threshold, interval, ratio, storage key, or any value used twice | | [docs/conventions/contributing-flow.md](docs/conventions/contributing-flow.md) | Opening, describing, or reviewing
entireio/cli/CLAUDE.md
protocol](docs/architecture/external-agent-protocol.md) | | Security, privacy, scanners | [Security and privacy](docs/security-and-privacy.md), [implementation contracts](docs/development/checkpoint-implementation.md) | | Review command | [Review architecture](docs/architecture/review-command.md) | | Release artifacts, install scripts, version-check origins | [Release origins](docs/architecture/release-origins.md) | | E2E tests
Steady4 mentions in 2 weeks
No other project uses it yet
umbraco/Umbraco-CMS/CLAUDE.md
dotnet format) - No build warnings - **Merge Strategy**: Squash and merge (via GitHub UI) - **Reviews**: Required from code owners #### PR Naming Convention Use the format: `Area: Description (closes #IssueID)` **Examples**: | Area
InternLM/xtuner/.claude/CLAUDE.md
requesting review. 3. Resolve all review comments before merging — do not dismiss without explanation. --- ## Code Review Standards ### What to Check 1. **Correctness**: Does the code do what it claims
Jeric-X/SyncClipboard/CLAUDE.md
operations; categories alone do not skip tests when running the full suite locally. ### Code Style ```bash # Check formatting (from src/ directory) dotnet format --verify-no-changes --severity info --no-restore
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.