How real projects brief Codex, Cursor and every other agent that reads AGENTS.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.
MicrosoftAGENTS.mdmicrosoft/WinUI-Gallery/AGENTS.md
Follow the repository architecture, build, accessibility, and coding guidance in .github/copilot-instructions.md. For the full sample-authoring guidance, including how little description text to use and where options and output go, see the gallery-control-samples skill in .github/skills/gallery-control-samples/SKILL.md.
MicrosoftAGENTS.mdmicrosoft/DirectXShaderCompiler/AGENTS.md
- Start from evidence. Identify the exact pass, hot function, and triggering operation before proposing a fix. Enumerate every limit, cutoff, cache, and heuristic on that path. - Check this…
No other project uses it yet
VercelAGENTS.mdvercel-labs/wterm/AGENTS.md
All apps and examples with dev servers use portless to avoid hardcoded ports. Portless assigns random ports and exposes each app via .localhost URLs. Naming convention: - Docs app: docs.wterm → docs.wterm.localhost - Examples: <name>.wterm → <name>.wterm.localhost When adding a new app or example that runs a dev server, wrap its dev script with portless <name>: Do not add --port flags — portless handles port assignment automatically. Do not add portless as a project dependency; it must be installed globally.
Steady8 mentions in 2 weeks
No other project uses it yet
VercelAGENTS.mdvercel-labs/wterm/examples/nextjs/AGENTS.md
This version has breaking changes — APIs, conventions, and file structure may all differ from your training data. Read the relevant guide in node_modules/next/dist/docs/ (resolved from this file's directory; in monorepos the next package may not be visible from the repo root) before writing any code. Heed deprecation notices. This block is written and re-added by next dev — verify at node_modules/next/dist/server/lib/generate-agent-files.js. Removing it from a diff only re-creates the uncommitted change; committing it with your work keeps the tree…
Steady8 mentions in 2 weeks
No other project uses it yet
ElasticAGENTS.mdelastic/elasticsearch-net/AGENTS.md
The repository builds with the standard .NET SDK CLI; the SDK version is pinned in global.json. - Build the client: dotnet build src/Elastic.Clients.Elasticsearch/Elastic.Clients.Elasticsearch.csproj - Build and test the request converter: dotnet test RequestConverter.sln (requires the wasm-tools workload: dotnet workload install wasm-tools) A legacy Bullseye build system (./build.sh, .\build.bat, driven by the F# project build/scripts/scripts.fsproj) still backs a few CI workflows. It is planned to be modernized or removed; do not use it for local development. There is no separate lint…
Steady3 mentions in 2 weeks
No other project uses it yet
GoogleAGENTS.mdgoogle/recaptcha/AGENTS.md
This file provides authoritative context, architectural invariants, branching policies, and verification requirements for AI coding agents and maintainers working on google/recaptcha.
Hugging FaceAGENTS.mdhuggingface/blog/agents.md
We are releasing Transformers Agents 2.0! ⇒ 🎁 On top of our existing agent type, we introduce two new agents that can iterate based on past observations to solve complex…
Hugging FaceAGENTS.mdhuggingface/blog/zh/agents.md
我们推出了 Transformers 智能体 2.0！ ⇒ 🎁 在现有智能体类型的基础上，我们新增了两种能够 根据历史观察解决复杂任务的智能体。 ⇒ 💡 我们致力于让代码 清晰、模块化，并确保最终提示和工具等通用属性透明化。 ⇒ 🤝 我们加入了 分享功能，以促进社区智能体的发展。 ⇒ 💪 全新的智能体框架性能卓越，使得 Llama-3-70B-Instruct 智能体在 GAIA 排行榜上超越了基于 GPT-4 的智能体！ 🚀 快去体验，力争在 GAIA 排行榜上名列前茅！ 大型语言模型…
AWSAGENTS.mdawslabs/agentcore-samples/01-features/02-host-your-agent/01-runtime/04-coding-agents/03-code-agents-competition-e2e/coding_agents/codex/AGENTS.md
You are a coding agent running on AWS Bedrock AgentCore. You fix bugs in GitHub repositories by reading issues, applying fixes, and submitting PRs. You have a gateway MCP server connected that provides GitHub tools. Use them directly — they are available as native tools. When given a prompt, act immediately: 1. Extract the repository owner, repository name, and issue number from the user's message. 2. Use the MCP tools to read the issue, fix the code, and submit a…
AWSAGENTS.mdawslabs/agentcore-samples/01-features/02-host-your-agent/01-runtime/04-coding-agents/03-code-agents-competition-e2e/coding_agents/open-code/AGENTS.md
You are a coding agent running on AWS Bedrock AgentCore. You fix bugs in GitHub repositories by reading issues, applying fixes, and submitting PRs. You have a gateway MCP server connected that provides GitHub tools. Use them directly — they are available as native tools. When given a prompt, act immediately: 1. Extract the repository owner, repository name, and issue number from the user's message. 2. Use the MCP tools to read the issue, fix the code, and submit a…
AWSAGENTS.mdawslabs/agentcore-samples/02-use-cases/02-workflow-automation-agents/event-driven-claims-agent/AGENTS.md
For humans: This file provides context for AI coding assistants (Kiro, Cursor, Claude Code, GitHub Copilot). For the human-readable documentation, see docs/, README.md, or docs/tutorial.md. This project is an event-driven insurance claims processor built on Amazon Bedrock AgentCore. It uses a dual-agent architecture (Claims Processor + Validation Agent) with cost-based model routing (Sonnet for reasoning, Haiku for validation) and a deterministic execution phase. Deploys as a single CloudFormation stack (AgentCore-ClaimsAgent-dev) via the AgentCore CLI. Important: AgentCore resources (Runtime, Gateway, Memory,…
AWSAGENTS.mdawslabs/agentcore-samples/02-use-cases/02-workflow-automation-agents/receipts-intelligent-document-processing-agent/AGENTS.md
Guidance for AI agents and contributors working in this sample.
MicrosoftAGENTS.mdmicrosoft/tgrep/AGENTS.md
A short guide for AI agents (and the humans who wire them up) that want to use tgrep as a fast search tool inside a repository. It complements the README, which documents every flag; this file covers the few things an agent has to get right. tgrep is ripgrep with a pre-built trigram index and an optional server. A search resolves in this order: 1. Server running for this tree: query it over TCP. Fastest. A file watcher keeps the…
Steady2 mentions in 2 weeks
No other project uses it yet
GoogleAGENTS.mdgoogle/ksp/AGENTS.md
Welcome! This file defines the mandatory workflow, engineering methodology, and project context for AI coding agents working in the Kotlin Symbol Processing (KSP) repository. You are working on a compiler plugin and symbol processing engine. There must be zero speculation, hand-waving, or hallucination. where `<CURREN
Steady9 mentions in 2 weeks
No other project uses it yet
Hugging FaceAGENTS.mdhuggingface/datatrove/AGENTS.md
DataTrove is a library to process, filter, and deduplicate text data at very large scale. It provides prebuilt pipeline blocks with a framework to add custom functionality. Pipelines are platform-agnostic, running locally, on Slurm, or on Ray clusters. See README.md for detailed documentation on terminology, pipeline blocks, executors, and usage examples. These rules apply to agents working for external contributors. Hugging Face org members, and PRs that only fix docs typos or broken links, are exempt from the approval rule.…
LangChainAGENTS.mdlangchain-ai/langgraphjs/AGENTS.md
When Corridor's analyzePlan tool is available, create a plan and use the tool to analyze it before generating or modifying code. Apply the resulting security guidance before writing code. LangGraph is consolidating as the low-level runtime for agents and workflows: graph execution, checkpointing, streaming, and interrupt/resume. Critical production systems and agent harnesses run on it, so correctness and stability come first. It is moving away from being a friendly SDK for building agents directly; agent development belongs one layer up,…
Steady4 mentions in 2 weeks
No other project uses it yet
LangChainAGENTS.mdlangchain-ai/langgraphjs/docs/docs/agents/agents.md
This guide shows you how to set up and use LangGraph's prebuilt, reusable components, which are designed to help you construct agentic systems quickly and reliably. Before you start this tutorial, ensure you have the following: If you haven't already, install LangGraph and LangChain: Use createReactAgent to instantiate an agent: Use initChatModel to configure an LLM with specific parameters, such as temperature: See the models page for more information on how to configure LLMs. Prompts instruct the LLM how to…
Steady4 mentions in 2 weeks
No other project uses it yet
NVIDIAAGENTS.mdNVIDIA/cuda-python/benchmarks/cuda_bindings/AGENTS.md
Read the README.md in this directory for more details about the benchmarks. When generating code verify that that the code is correct based on the source for cuda-bindings that can be found in ../../cuda_bindings_13
NVIDIAAGENTS.mdNVIDIA/cuda-python/benchmarks/cuda_core/AGENTS.md
Read the README.md in this directory for more details about the benchmarks. When generating code verify that that the code is correct based on the source for cuda-core that can be found in ../../cuda_core. This suite shares the pyperf runner with ../cuda_bindings/ via a sys.path insert in run_pyperf.py. The per-suite setup (runtime.py, the benchmarks/ module files) lives here. Benchmark IDs are kept identical to the cuda.bindings suite so compare.py can diff them directly.
NVIDIAAGENTS.mdNVIDIA/cuda-python/cuda_bindings_12/AGENTS.md
This root maintains CUDA 12.9 bindings under the cuda.bindings namespace. Follow the repository-wide instructions, including Changes across bindings majors. Inspect the CUDA 13 counterpart during implementation and review; preserve intentional API, layout, and build differences. - Files carrying a CYTHON-BINDINGS-GENERATED seal are generated inputs. Change their generation source and regenerate against CUDA 12.9 inputs. Record generator revision, toolkit inputs, and affected files. Do not edit sealed files or recalculate their seals to conceal manual changes. - Handwritten utilities, build helpers,…
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.
