NVIDIAKnown publisherNVIDIA/cloudai/AGENTS.md

cloudai

AGENTS.md103 starsChanged 58 days agoApache-2.0View on GitHub

- Avoid over-engineering - Prefer Google Python style guide: - Blend into existing code - No asserts in production code - Prefer absolute imports (import x). Use from x import y when qualified names are too long. Use relative imports when existing code uses it. - Backwards compatibility is very important. CloudAI may be integrated into other tools. Users maintain their own CloudAI configs that we may never see. Backwards incompatible changes should be avoided unless explicitly asked. In that…

Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,700 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository8 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- cloudai-config-contributionSkill
- cloudai-docs-contributionSkill
- cloudai-remote-experimentSkill
- cloudai-remote-installSkill
- cloudai-workload-contributionSkill
- pr-managementSkill
- security-assessmentSkill

The file

Show its rules and text

We split this AGENTS.md into the rules it gives an agent

14 rules. About 1 is common, 4 are distinctive, and 9 are very specific to this project.

≈1
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
9
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific
could be automatic: a setting could enforce it

Distinctive:
Prefer Google Python style guide:
line 6
Distinctive:
No asserts in production code
line 8
Very specific:
Prefer absolute imports (import x).
line 9
Very specific:
Use from x import y when qualified names are too long.
line 10
Distinctive:
Use relative imports when existing code uses it.
line 11
Very specific:
Update affected documentation in README.md or doc/ when public behavior changes.
line 16
Very specific:
Import public core APIs through cloudai.core, respect import-linter boundaries, and use existing lazy-import mechanisms for heavy modules.
line 17
Very specific:
Follow CONTRIBUTING.md, including SPDX headers and mirrored tests for new Python modules.
line 19
Very specific:
Start with focused tests: uv run --locked --extra dev pytest <test-paths>.
line 23
Very specific:
Run uv run --locked --extra dev pre-commit run --files <changed-files> and review formatter edits.
line 24
Distinctive:
Do not commit, push, open or modify pull requests, or run remote jobs unless explicitly asked to do so.
line 28
Very specific:
Follow PR requirements in CONTRIBUTING.md and use .github/PULL_REQUEST_TEMPLATE.md.
line 29
Very specific:
Use .agents/skills/pr-management/SKILL.md for PR creation/updates and after every push to a branch with an open PR.
line 30
Common:
Keep public configs reusable (no internal resources).
line 31

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

AGENTS.md

CloudAI repository guidance

Implementation conventions

Avoid over-engineering
Prefer Google Python style guide:
- Blend into existing code
- No asserts in production code
- Prefer absolute imports (import x). Use from x import y when qualified names are too long. Use relative imports when existing code uses it.
Backwards compatibility is very important. CloudAI may be integrated into other tools. Users maintain their own CloudAI configs that we may never see. Backwards incompatible changes should be avoided unless explicitly asked. In that case the changes must be highlighted.
Update affected documentation in README.md or doc/ when public behavior changes.
Import public core APIs through cloudai.core, respect import-linter boundaries, and use existing lazy-import mechanisms for heavy modules. Follow established workload structure and registration patterns.
Follow CONTRIBUTING.md, including SPDX headers and mirrored tests for new Python modules.

Verification

Start with focused tests: uv run --locked --extra dev pytest <test-paths>.
Run uv run --locked --extra dev pre-commit run --files <changed-files> and review formatter edits.

Contribution

Do not commit, push, open or modify pull requests, or run remote jobs unless explicitly asked to do so.
Follow PR requirements in CONTRIBUTING.md and use .github/PULL_REQUEST_TEMPLATE.md.
Use .agents/skills/pr-management/SKILL.md for PR creation/updates and after every push to a branch with an open PR.
Keep public configs reusable (no internal resources). Never commit credentials, private artifacts, or internal hostnames.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

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.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free