How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
joshzyjCLAUDE.mdjoshzyj/open-scholar-skill/CLAUDE.md
Trace sidecar contract. ### Output Formats Manuscript skills generate 4 formats: `.md`, `.docx`, `.tex`, `.pdf` via pandoc. --- ## Reference Manager Integration Unified search layer: `.claude/skills/_shared/refmanager-backends.md` | Tier | Backend | Notes | |------|---------|-------| | 0 | Knowledge Graph | `scholar
ever-worksCLAUDE.mdever-works/ever-works/CLAUDE.md
perplexity, brightdata, firecrawl, jina, valyu, linkup - **content-extractor**: local-content-extractor (default), notion-extractor, pdf-extractor, scrapfly - **screenshot**: screenshotone, urlbox, scrapfly - **git-provider**: github (default + OAuth) - **deployment**: vercel (default) - **data
oaustegardCLAUDE.mdoaustegard/claude-skills/CLAUDE.md
verb+ing form): - ✅ CORRECT: `creating-mcp-servers`, `processing-pdfs`, `updating-knowledge` - ❌ WRONG: `mcp-creator`, `pdf-processor`, `knowledge-update` These are non-negotiable requirements for all skills in this repository. ### CLAUDE.md
SerhiiKorniienkoCLAUDE.mdSerhiiKorniienko/bullshit-detector/CLAUDE.md
captions), an article, a tweet (`https://x.com/naval/status/1002103360646823936` works), and a PDF. The `agents/` directory ships with the Claude Code plugin (auto-discovered) and holds subagents that skills delegate
AnkitClassicVisionCLAUDE.mdAnkitClassicVision/Claude-Code-Deep-Research/CLAUDE.md
Tools to use: 1. WebSearch for academic papers and technical resources 2. WebFetch for PDF extraction and content analysis 3. mcp__filesystem__ tools to save important findings Look for: - Peer
DjangoPengCLAUDE.mdDjangoPeng/agentic-ai/ai-quant-cli/CLAUDE.md
repository. > 本文件只管 `ai-quant-cli/` 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 从年报 PDF 造一个本地系统：解析合并三张报表 → **由 Claude Code 亲自做财务风险研判** → 出图 → 生成 HTML 投研报告。构建按 `BUILD_PLAN.md` 的 6 个阶段推进
DjangoPengCLAUDE.mdDjangoPeng/agentic-ai/CLAUDE.md
uploaded attachments into a job workspace. 2. Ingest supported PDF/image documents. 3. Extract native PDF text or OCR text, then parse invoice fields. 4. Validate required fields, confidence, compliance, deduplication
aker-devCLAUDE.mdaker-dev/microfolio/CLAUDE.md
local one carry the same images. `example_projects.zip` ships only JPEGs plus one small real PDF per project, and the demo was published at full size until that was true
AgriciDanielCLAUDE.mdAgriciDaniel/claude-repurpose/CLAUDE.md
knowledge files repurpose-twitter/SKILL.md # Twitter/X (threads, tweets, polls) repurpose-linkedin/SKILL.md # LinkedIn (posts, PDF carousels, polls) repurpose-instagram/SKILL.md # Instagram (carousels, captions, reels) repurpose-tiktok/SKILL.md # TikTok (video scripts
AWSCLAUDE.mdawslabs/visual-asset-management-system/web/src/visualizerPlugin/CLAUDE.md
flac, .m4a | enabled | | `columnar-viewer` | Columnar Data Viewer | data | .fcs, .csv | enabled | | `pdf-viewer` | PDF Viewer | document | .pdf | enabled | | `cesium-viewer` | Cesium 3D Tileset | 3d | .json | enabled | | `text-viewer` | Text
sweetrbCLAUDE.mdsweetrb/apple-notes-mcp/CLAUDE.md
title:`, `body:`, `text:`, `folder:`, `account:`, `tag:`; `has:link|attachment|checklist|drawing|image|video|audio|pdf|table|scan|tag`; `checklist:open|done`; `pinned`, `locked`, `shared`; `words:>250`; `created
JasonColapietroCLAUDE.mdJasonColapietro/suede-creator-skills/CLAUDE.md
every stale count from the live measurement npm run build:book # regenerates BOOK.md, the PDF and docs/book from those sources npm run build:cards # regenerates the per-skill CARD.md release
kv0906CLAUDE.mdkv0906/pm-kit/CLAUDE.md
read/write/link operations | ### Export Flags Skills that generate output support format export flags: `--xlsx`, `--docx`, `--pdf`, `--pptx`. Append to any supported skill invocation (e.g., `/progress all --xlsx`, `/doc project
sero12journeyCLAUDE.mdsero12journey/synopsis-to-scenario/CLAUDE.md
real stories) into Korean feature-length film screenplays. Final output: completed Korean screenplay (.docx / .pdf). ## Architecture Four agents, strict separation of concerns: | Agent | Role | Steps | |-------|------|-------| | **Orchestrator** | Workflow control, data passing
daypunkCLAUDE.mddaypunk/LockedIn/CLAUDE.md
edge predicates, v3 ├── storage/{notes,graph}.py ← markdown read/write, edge derivation ├── ingest/{interview,markdown,pdf,docx,text}.py ├── render/_template.py ├── hud/oauth_usage.py ← Anthropic OAuth credential reader + caller └── commands/{install,doctor,validate,init
NeoTheCaptCLAUDE.mdNeoTheCapt/RedteamAgent/agent/CLAUDE.md
Board visit and fresh `/api/Challenges` fetch. - `Upload Type`: after `/file-upload` acceptance, submit a non-PDF/non-ZIP payload and check the consumer/scoreboard result instead of stopping at ZIP acceptance. - `User Credentials
songzhuozhuCLAUDE.mdsongzhuozhu/obsidian-llm-wiki/inspool-wiki-en/CLAUDE.md
Stores raw source material such as web clips, articles, book excerpts, meeting notes, PDF transcripts, screenshots, and similar inputs. - This is the factual source layer. - The body is read-only
songzhuozhuCLAUDE.mdsongzhuozhu/obsidian-llm-wiki/inspool-wiki-zh/CLAUDE.md
原始资料可以保留原语言，不需要强制翻译；但 wiki 沉淀层的说明和结构应默认使用中文。 ## 三层结构 ### 1. `raw/` - 存放原始资料，例如网页剪藏、文章、书摘、会议纪要、PDF 转写稿、截图说明等。 - 这是事实来源层。 - 正文默认只读。 - 默认允许两类流程性变更：移动文件路径，以及更新少量 frontmatter 处理元数据。 - 除非用户明确要求，否则不得改写、删减或重写 `raw/` 内正文内容。 ### 2. `wiki
Wide-MoatCLAUDE.mdWide-Moat/open-computer-use/CLAUDE.md
mount point) to avoid duplication per container: | Path | Contents | Storage | |------|----------|---------| | `/home/node_modules/` | Libraries (react, pptxgenjs, pdf-lib...) | Image layer (shared) | | `/usr/local/lib/node_modules_global/` | CLI tools (mmdc, tsc, tsx, claude) | Image layer (shared) | | `/home/assistant/node_modules
discourselabCLAUDE.mddiscourselab/scrapai-cli/CLAUDE.md
when they hold any content. When unsure whether something is content → **include it.** **PDF links are collected by default** (recorded as URL items — `PDF_MODE: links_only`); set `PDF_MODE
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.
