How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
Hugging FaceAGENTS.mdhuggingface/lerobot-humanoid-design/AGENTS.md
This repository is one part of the LeRobot humanoid stack: Any change here should preserve compatibility with the other four repositories. Develop and validate mechanical/control design choices before hardware commits:
Hugging FaceAGENTS.mdhuggingface/lerobot-humanoid-identification/AGENTS.md
This repository is the identification block in the LeRobot humanoid stack: Identification outputs are used to tune simulator realism for runtime and training loops. Run joint-wise simulator identification with MJWarp + CMA-ES: Run at least one smoke identification command and verify: Preferred smoke shape: CPU, small popsize, maxiter=1, single joint.
Hugging FaceAGENTS.mdhuggingface/agentcap/AGENTS.md
Read this before changing the code. The README explains what the project is and how to use it; this file holds decisions that already hardened and what's not yet built. These are settled. If you want to revisit any, raise it with the user explicitly first. 1. Capture is dumb. The proxy persists raw OpenAI-compat bytes. No tokenisation, no chat-template render at capture time. The capture path must work without the model's tokenizer being loaded. 2. Export is a dumb…
Hugging FaceAGENTS.mdhuggingface/transformers-mlinter/.ai/AGENTS.md
Run make format or make lint and make test before wrapping up a change. - Rule modules live in mlinter/trf*.py. - Rule metadata lives in mlinter/rules.toml. - Rule-specific tests live in tests/testtrfXXX.py, one file per rule. - Shared rule-test helpers live in tests/ruletest_utils.py; rule test classes should inherit from RuleTestCase when they use the standard run(rule, source, filename=...) helper. - General CLI, registry, cache, discovery, and public-API tests live in tests/test_mlinter.py. - Skills live under .ai/skills/. - The docs site…
Hugging FaceAGENTS.mdhuggingface/research-agent/AGENTS.md
Reference repositories: fast-agent: ~/source/fast-agent/ prefect fastmcp: ~/source/fastmcp/ model context protocol: ~/ Use codexresponses.gpt-5.6-terra?reasoning=high subagents for research, analysis, review and code production. Design: - Prefab source: research/appui.py; preview states: design/previewstates.py. -…
Hugging FaceAGENTS.mdhuggingface/swarm-sweeper/AGENTS.md
Keep imports directional: Enforce this with: This check belongs in CI and in the local pre-commit/pre-push validation flow. When inspecting this repository, never search outside the workspace root unless the…
Hugging FaceAGENTS.mdhuggingface/new-model-addition-cwm/AGENTS.md
This AGENTS.md file provides guidance for code agents working with this codebase. Many models in the codebase have similar code, but it is not shared by inheritance because we want each model file to be self-contained. We use two mechanisms to keep this code in sync: - "Copied from" syntax. Functions or entire classes can have a comment at the top like this: # Copied from transformers.models.llama.modelingllama.rotatehalf or # Copied from transformers.models.t5.modelingt5.T5LayerNorm with T5->MT5 These comments are actively checked by…
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.
