pr-workflow-vp

PR Writer

Use this skill for external open-source PR work. The workflow has six phases:

1. Step1 Fork & Clone
2. Step2 Find PR Chances
3. Step3 Validate PR Value
4. Step4 Implement & Create Formal PR
5. Step5 Address Reviewer Feedback
6. Step6 Refine PR Comment Section

Non-Negotiables

• Always start multi-agent work when this skill runs. Launch at least two agents before phase-specific work; use more for Step2, Step3, Step4, Step5, and Step6 when the repository is large, the task is ambiguous, reviewer feedback touches behavior-sensitive code, or a PR comment section needs stronger evidence.
• Treat submission as dry-run by default outside Step4 and Step5. Do not commit, push, create, mark ready, or update a PR unless the user explicitly asks for that action. When the user invokes Step4 by asking to create or submit a PR, treat that as permission to commit, push, and create a formal ready-for-review PR by default, unless the user explicitly asks for a draft PR.
• Do not modify references/ files as part of ordinary PR workflow.
• Default clone path is C:\XingyueZhu\Github\repos\<repo-name> on this Windows machine unless the user gives another path.
• Prefer small, reviewable bug fixes. The default target is a production-code fix of 0-20 changed lines.
• Do not include unrelated refactors, generated churn, broad formatting, or CHANGELOG.md unless the project explicitly requires it.
• Keep evidence honest. Distinguish runtime reproduction, UI reproduction, package reproduction, focused unit reproduction, inferred risk, and unavailable validation.
• Never add Co-authored-by trailers. Do not add bot co-authors, and do not add human reviewers as co-authors either. Commit authorship stays with the actual committer only. (User instruction, 2026-10-29: "Co-authored-by 目前不要去加这一段".)
• Read references/pr-examples.md before drafting or revising PR prose. Treat it as the few-shot style guide for concise, evidence-driven bug-fix PRs.
• PR delivery follow-ups (Step4 and Step5). After a PR is created or updated, do two things before reporting. Both are environment-dependent: ask once, then reuse the answer.
  1. Open the PR in whatever browser surface the current IDE actually provides. The IDE may be Qoder, WorkBuddy, Codex, or something else — do not hardcode a product or a tool name. Use the in-app browser/preview panel when the runtime exposes one; if not, report that the surface is unavailable and hand the raw URL to the user. Ask the user on the first run of a new IDE which surface they want.
  2. Produce a Chinese-translation Markdown copy of the PR body. Confirm the output location with the user on first use instead of assuming a fixed path; keep it outside the git work tree so it never pollutes the diff, and preserve the PR template headings, checkboxes, and status. The English body and the GitHub URL stay authoritative.
• Do not loop on summaries. When a phase completes and the only remaining work is waiting on an external event (maintainer review, CI, merge), report the final state once and hand control back to the user instead of re-printing the same summary on every "continue". Mark wait-on-review phases as pending (or in_progress with a "waiting for reviewer" note), never completed; a phase is only complete when its deliverable exists. This is a default, not a hard rule: if the user asks to keep going, keep going.