Repository navigation
docs: add contribution guide and AI-assisted contribution policy - #1354
Merged
Merged
Conversation
- CONTRIBUTING.md: PRs no longer need an issue first (larger changes still should), at most 3 open PRs per external contributor, scope rule for undocumented parameter combinations, AI policy (human owns it, discloses it, answers reviews), and what maintainers do when it's not followed. - AGENTS.md: instructions for AI agents to re-evaluate and close low-value PRs/issues. - PR and issue templates: required AI-assistance disclosure; PR template drops the issue-first requirement. - README: point to CONTRIBUTING.md instead of the issue-first callout. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… templates Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ed path Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Blank issues bypassed the required AI-assistance dropdown in the issue forms. The PR template asked for a description "below" with nowhere to write it, and didn't say the checkboxes are exclusive. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Mention the CLA in the pull request section. - Broaden "answer reviews yourself" to "speak for yourself": allow AI for translation and grammar, and ask for a human summary at the top of AI-written PR descriptions (also in AGENTS.md and the PR template). - Exempt automation set up by the maintainers, such as Dependabot. - Require two maintainers to agree on a block, add an appeal path, and add a standard closing message. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Disclosure only asks whether AI was used, not how. - The docs-improvement form called the issue a "report"; say suggestion. - The standard closing message links the policy without a placeholder. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Collaborator
Author
|
@eddiebergman-priorlabs looks great, thank you for changes! Can you approve the PR then? Since it's mine I can't |
eddiebergman-priorlabs
approved these changes
Oct 12, 2026
eddiebergman-priorlabs
enabled auto-merge
October 12, 2026 17:35
oscarkey
approved these changes
Oct 13, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Issue
Internal: OPC-28 (AI contribution policy). Follows the #engineering thread on unattended agent PRs/issues.
Motivation and Context
We're getting bulk agent-generated issues and PRs with no human answering reviews (e.g. #1344–#1353, opened within ~30 seconds this morning). Some of the earlier ones were genuinely useful and got merged, so this doesn't ban AI. It gives maintainers a written policy they can link to when closing, in line with LLVM and SciPy, plus a vLLM-style
AGENTS.md.What changes:
CONTRIBUTING.md(new): AI tools are fine; a human owns the contribution, discloses AI use and answers reviews themselves. No fully automated or bulk submissions. Max 3 open PRs per external contributor. For undocumented parameter combinations we generally prefer a clear error over new code paths. Maintainers may close non-compliant PRs/issues without detailed review; repeat violations can lead to a block.AGENTS.md(new): tells agents to re-evaluate their PR againstCONTRIBUTING.mdand close it if it isn't worth the review time.CONTRIBUTING.md.Open for discussion: the PR cap (3), and whether we want a cap on open issues too.
Public API Changes
How Has This Been Tested?
Docs and templates only. Issue-template YAML parses.
AI Assistance
Checklist
🤖 Generated with Claude Code