AGENTS.md examples from real projects

Google DeepMindAGENTS.md

acme / user

google-deepmind/acme/docs/user/agents.md

Acme includes a number of pre-built agents listed below. All agents can be used to run synchronous single-threaded or distributed experiments. Distributed experiments are using Launchpad and can be executed either on a single machine (--lp_launch_type=[local_mt|local_mp] command line flag for multi-threaded or multi-process execution) or multi machine setup on GCP (--lp_launch_type=vertex_ai). For details please refer to Launchpad documentation. We've listed the agents below in separate sections based on their different use cases, however these distinction are often subtle. For more…

Google DeepMindAGENTS.md

mujoco_playground

google-deepmind/mujoco_playground/AGENTS.md

- Always use uv run, not bare python. Pass explicit --extra flags (e.g. --extra rl, --extra so101, --extra yam) to uv run rather than --no-sync; pyproject.toml (Python 3.12–3.13) configures PyPI as the default index, CMAKE_CXX_STANDARD=17 for reaf, and cross-platform override-dependencies (evdev on Linux only, ruckig>=0.15.3). - Always use absl (from absl import app, flags, logging) for scripts, tools, and entrypoints. NEVER use argparse. - Dependencies are managed with uv. Default make sync installs --extra rl --group dev with CUDA 13…

No other project uses it yet

Google DeepMindAGENTS.md

mujoco_warp

google-deepmind/mujoco_warp/AGENTS.md

See README.md for development setup. @CONTRIBUTING.md - PR body should be plain, concise prose. Describe the problem, what the change does, and any non-obvious tradeoffs. Bullet points listing changes are fine, but avoid section headers, structured templates, and emojis. A good PR description reads like a short paragraph to a colleague, not a form. - PR and commit messages are rendered on GitHub, so don't hard-wrap them at 88 columns. Let each sentence flow on one line.

No other project uses it yet

Google DeepMindAGENTS.md

formal-conjectures

google-deepmind/formal-conjectures/AGENTS.md

Formal Conjectures states open mathematical problems in Lean 4. It is a statement repository, not a proof repository: almost every problem is sorry. CONTRIBUTING.md is the reference for conventions, folders, and attributes. Also follow any README.md in the directory that you change. Use these detailed guides when the task needs them: Follow the style of an existing file. Start by copying one from the same directory. Search Mathlib, FormalConjecturesForMathlib/, and nearby problem files before you add a definition, API, or…

No other project uses it yet

Google DeepMindAGENTS.md

simply

google-deepmind/simply/AGENTS.md

Simply is a minimal JAX-based research codebase for LLM training and inference. It emphasizes minimal abstractions for rapid iteration on frontier research. The codebase supports Gemma, Qwen, and DeepSeek model families with multi-host distributed training.

Google DeepMindAGENTS.md

agentic_ecology / assets

google-deepmind/agentic_ecology/skills/agentic-ecology-init/assets/AGENTS.md

- Environment Management: This repository is managed exclusively with uv. All dependency resolution and environment setup are handled by uv. The soundfile library is guaranteed to be available in this environment; do not waste steps verifying its installation. - Agent Artifacts: You MUST store artifacts (Python scripts, Markdown documents, etc.) in the agent_workspace directory inside the repository's root directory so that the user can inspect your work. Do NOT create or commit any temporary, throwaway, or scratch files anywhere else.…

Google DeepMindAGENTS.md

padic-ml

google-deepmind/padic-ml/AGENTS.md

Precision Convention: - We should not hardcode precision. Use jnp.int_ or jnp.float_. User scripts should control precision via globally setting jax_enable_x64 themselves.