How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
charmonyeAGENTS.mdcharmonye/landkit/AGENTS.md
without scripts or mandatory commands. - `photo-intake-processing` - review, select, and prepare site photos. - `pdf-doc-processing` - read and extract useful information from PDF documents. - `ai-concept-generation` - prepare
hletrdAGENTS.mdhletrd/kiwi-paper/AGENTS.md
괄호 코멘터리**: (이런 식으로 중간에 편집자 의견을 삽입) ### 4단계 파이프라인 1. **초안 작성 (Draft)**: PDF 내용을 나무위키 스타일로 변환 2. **다듬기 (Refine)**: 구조, 유머 밸런스, 링크를 보강 3. **휴머나이즈 (Humanize
vaibhavpandeyvpzAGENTS.mdvaibhavpandeyvpz/hooman/AGENTS.md
review"`) → human review → ask export format → delivery export → then `stop_design_preview`. Delivery formats: `pdf` / `images-to-pdf` / PowerPoint-ready `pptx` / Figma-ready `figma` / `figma-deck` / Sketch-ready `sketch
quan0715AGENTS.mdquan0715/open-press/AGENTS.md
slide` Press for multi-Press verification. Use it to validate real content, style, PDF, deploy, and gallery routing. - `docs/` — user-facing docs, migration notes, active specs, and implementation plans. - Root
axisroboAGENTS.mdaxisrobo/ea-harness/AGENTS.md
vision) → partial req.yaml | | arch-req-from-doc | `/arch-req-from-doc` | `@arch-req-from-doc` | Reader — PDF / DOCX / MD / TXT via LLM → partial req.yaml | | arch-req-from-api | `/arch-req-from-api` | `@arch-req-from
aytzeyAGENTS.mdaytzey/paper-pilot/AGENTS.md
search across 6 databases (Semantic Scholar, OpenAlex, arXiv, Crossref, Europe PMC, DOAJ) - open-access PDF resolution and inspection - Sci-Hub paper resolution and download (opt-in, disabled by default) - deep
dirmacsAGENTS.mddirmacs/daedra/AGENTS.md
classifier: genuine no-results or soft block | | `src/tools/fetch.rs` | `FetchClient`, `FetchedContent` (Html/Pdf/Document/Text/Binary), dom_smoothie, infer, pdf-inspector, anydoc | | `src/tools/crawl.rs` | Site crawl (sitemap + link following) | | `src/tools/bing.rs`, `src/tools/google.rs` | HTML scraper backends | | `src/tools
siropkinAGENTS.mdsiropkin/chrome-bridge/AGENTS.md
command whose fresh mark hits a page Chrome forbids scripting (chrome://, Web Store, PDF, another extension's page) releases that mark again itself — the error says so, and no pill
vin-spiegelAGENTS.mdvin-spiegel/htmldrop/AGENTS.md
image/*` (png, jpeg, gif, webp, svg). Text formats render into a reader page; PDF and images are stored and served verbatim. Pass a title via the `x-htmldrop-title` header
ahundtAGENTS.mdahundt/autorun/AGENTS.md
execution with three-stage verification, file policies, safety guards, task tracking, plan export) and **pdf-extractor** (`/pdf-extractor:extract`, `extract-pdfs`). `gemini` in code and docs means the Qwen Code / Antigravity
LeeYudokAGENTS.mdLeeYudok/doksam-skills/AGENTS.md
배로 축소되어 인라인 `top` 만으로는 정렬을 알 수 없다. - `skills/mobile-web-planner/scripts/export_deck.py`: 산출물 HTML 에서 PDF 와 PPTX 를 함께 만드는 내보내기 스크립트. - `skills/mobile-web-planner/scripts/check_layout_runtime.py`: Chrome headless 로 렌더해 레이아웃 회귀(슬라이드 overflow
MigchwAGENTS.mdMigchw/llm-wiki-starter/.agents/AGENTS.md
Fallback):** หากเจอเว็บที่เป็น Dynamic/SPA (React, Vue, SET) สลับไปใช้ Headless Chromium อัตโนมัติ - **Tier 3 (MarkItDown):** สำหรับแปลง PDF, PPTX, DOCX, XLSX ในเครื่อง - **Tier 4 (Video & Keyframe Capture):** สำหรับ YouTube / วิดีโอ ดึง transcript พร้อมสกัดภาพแผนภูมิ
urmzdAGENTS.mdurmzd/incipit/AGENTS.md
Overview **incipit** is a Go CLI tool that converts structured resume data (JSON/Markdown) into PDF, HTML, LaTeX, DOCX, and Markdown output formats. It includes `ai` commands that print self-contained
a-attiaAGENTS.mda-attia/scicomp-research-skills/AGENTS.md
drafting, paragraph-clarity check, claim-evidence alignment, adversarial review. | upstream (Master-cai) | | `skills/literature-survey/` | bibtex + PDF + pdftotext + per-paper survey-note + collection-log workflow for heavy-literature papers. | added here | | `skills/human-facing-doc-authoring
docxologyAGENTS.mddocxology/codomyrmex/AGENTS.md
Ollama) - Key Classes: `VLMClient`, `PDFExtractor`, `AnnotationExtractor` - Key Functions: `analyze_image(path) -> VLMResponse`, `extract_text(pdf_path) -> str` - `bio_simulation/` - Biological simulation - Key Classes: `BioSimulator` - `crypto/` - Cryptographic utilities - Key Classes: `CryptoEngine
GigleAIAGENTS.mdGigleAI/cavil-loop/AGENTS.md
hook) │ ├── session-log.sh ← View tmux pane history │ └── weekly-report/ ← Monday auto weekly report (collect / render / PDF / publish) ├── prompts/ │ ├── new-issue.template.md ← Prompt for new-issue dispatch │ ├── issue-comment.template.md ← Prompt for new issue comment │ └── pr-comment.template.md
NemoArce2007AGENTS.mdNemoArce2007/cumcm-skills/AGENTS.md
skill 才能再调模型技能；用户技能之间禁止互调。 5. 验收失败不得把 `contest-state.json` 的 `status` 前移。 6. 中文回复用户。技能文件本身也用中文，专有名词保留 CUMCM、PDF、AHP 等。 ## 技能怎么选 不确定 → 读本文件下面的表，或让用户跑 `/ask-cumcm`。 | 用户说了什么 | 用 | 不要用 | | --- | --- | --- | | 第一次在本仓库/赛题目录用这套技能 | `setup-cumcm-skills
rockbenbenAGENTS.mdrockbenben/ai-job-search-cn/AGENTS.md
还是占位符）。输出三段：环境哪几项就绪、该用户的流水线走到哪、 **下一步该做什么（只给一条）**。 - **把「下一步」那条直接转述给用户**，不要让他自己解读输出。 - 环境缺项**只在与用户当前意图相关时**才提（他要出 PDF 而 typst 缺失 → 说； 否则别把七行检查全念一遍）。 - **缺依赖不是拒绝理由**：只影响对应命令，其它照常，各工作流自带降级路径。 - 用户已明确说了要做什么、且与环境无关（如「改个措辞」）→ 跳过自检，别变成仪式
alejooroncoyAGENTS.mdalejooroncoy/campus-cli/AGENTS.md
Download a bbcswebdav URL directly | | `blackboard_upload_attempt_file` | Upload a local file (image, PDF, etc.) and get a fileUploadId | | `blackboard_save_attempt_draft` | Save text/files to an attempt WITHOUT
bks-labAGENTS.mdbks-lab/open-bridge/AGENTS.md
spacing, component anatomy). **Read it before generating user-facing visuals** — any skill emitting HTML, PDF, slides, certificates or styled email pulls its tokens from there. **Never hand-pick a brand
An open format for instructions to coding agents, read by Codex, Cursor and others. Think of it as a README written for agents.
At the repository root, with more specific files in subdirectories. Agents read the one closest to the file they're editing.
Setup and test commands, code style, and the rules a new contributor would need to know.
Claude Code reads CLAUDE.md. A one-line CLAUDE.md that points at AGENTS.md covers both.
