agent doc: add a requirement for PR communications - #14709
Merged
Merged
Conversation
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HCdbvRpL9fv3h3WwSUPpfS
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HCdbvRpL9fv3h3WwSUPpfS
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HCdbvRpL9fv3h3WwSUPpfS
…d human-readable Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01HCdbvRpL9fv3h3WwSUPpfS
github-actions
Bot
added
the
documentation
yiyixuxu
changed the title
modular.md helpers take explicit arguments, not components
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
stevhliu
approved these changes
Sep 7, 2026
Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>
yiyixuxu
commented
Sep 7, 2026
7 tasks
NancyFyong
added a commit
to NancyFyong/verl-omni
that referenced
this pull request
Sep 23, 2026
AGENTS.md required the human submitter to review every changed line, but said nothing about the text around the code: commit messages, the PR description and GitHub replies, or when an agent may publish them. diffusers added a similar rule in huggingface/diffusers#14709. AGENTS.md now requires that text to be written or approved word for word by the submitter, and lists the publishing actions an agent takes only when asked. Unlike diffusers, an agent may still post when the user asks for that specific action, which is what self-review already allowed. CONTRIBUTING.md tells contributors they own what they post, and self-review links to AGENTS.md instead of keeping its own list. AI assistance (Claude Code) was used for this change. Co-authored-by: Claude Code <noreply@anthropic.com> Signed-off-by: NancyFyong <88076188+NancyFyong@users.noreply.github.com>
NancyFyong
added a commit
to verl-project/verl-omni
that referenced
this pull request
Sep 24, 2026
…an-in-the-loop rule (#665) * [doc] chore: refine agent code-style rules and self-review skill The agent code-style rule mixed automated gates with review conventions and overstated what some hooks enforce (mypy runs with global ignore_errors; check-license only matches copyright text), so agents treated a green hook as proof of coverage. Split the two explicitly and point to the config each gate reads instead of snapshotting it. Add the conventions reviewers already apply but no rule wrote down, reconciled with the review template in zhaochenyang20/sglang-diffusion-routing#32 and with verl/diffusers practice: runtime-boundary validation and state ownership, example shell recipes, signatures and closed-set dispatch, descriptive local names, and inlining single-caller helpers. Replace the fixed comment density figure and copy-line threshold with contextual guidance. pipelines.md required lazy diffusers/vllm_omni imports, but every adapter imports them at module scope and the CPU job installs both, so defer to code-style's feature-boundary rule instead. config.md now spells out the deprecate-then-remove sequence #522 followed. Restructure the self-review skill into purpose, code quality, goal completeness and validation, with severity separate from category, evidence tags, and no posting or CI triggering without authorization. List the self-review and profile skills in the .agents index. AI assistance (Claude Code) was used for this change. Co-authored-by: Claude Code <noreply@anthropic.com> Signed-off-by: NancyFyong <88076188+NancyFyong@users.noreply.github.com> * [doc] chore: require human approval for reviewer-facing text AGENTS.md required the human submitter to review every changed line, but said nothing about the text around the code: commit messages, the PR description and GitHub replies, or when an agent may publish them. diffusers added a similar rule in huggingface/diffusers#14709. AGENTS.md now requires that text to be written or approved word for word by the submitter, and lists the publishing actions an agent takes only when asked. Unlike diffusers, an agent may still post when the user asks for that specific action, which is what self-review already allowed. CONTRIBUTING.md tells contributors they own what they post, and self-review links to AGENTS.md instead of keeping its own list. AI assistance (Claude Code) was used for this change. Co-authored-by: Claude Code <noreply@anthropic.com> Signed-off-by: NancyFyong <88076188+NancyFyong@users.noreply.github.com> --------- Signed-off-by: NancyFyong <88076188+NancyFyong@users.noreply.github.com> Co-authored-by: Claude Code <noreply@anthropic.com>
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
4 participants
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.
No description provided.