Menu
End-to-end VBW issue fix workflow using Codex planner, QA agents, worktrees, tests, and PR/CI gates. Explicit invocation only.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Use when explicitly asked to review a contributor or external PR for VBW against the linked issue, an independent fix plan, local tests, QA evidence, and repo conventions. Submit a real GitHub PR review with inline code comments for anchorable findings; explicit invocation only.
Shared VBW full-contract QA rubric for Codex QA agents reviewing issue fixes and contributor PRs.
This skill should be used when the user asks to run or verify a VBW smoke test with plain Claude Code (`claude`), capture visible `Skills:` output, validate `.vbw-planning/.skill-decisions.log`, or confirm consumer-smoke-worktree versus `--plugin-dir` provenance for local smoke coverage in this repo.
Use this skill when asked to run or verify a VBW smoke test through claude-code-router (`ccr code`), check visible `Skills:` output, validate `.vbw-planning/.skill-decisions.log`, or confirm consumer-smoke-worktree versus `--plugin-dir` provenance.
Based on SOC occupation classification
| name | vbw-fix-issue |
| description | End-to-end VBW issue fix workflow using Codex planner, QA agents, worktrees, tests, and PR/CI gates. Explicit invocation only. |
Use this skill when the user explicitly asks to fix a VBW issue, implement a tracked GitHub issue, or run the repo issue-fix workflow.
Do not use this skill for casual code review, one-off explanations, or untracked refactors.
Run these phases in order. Keep every shell command rooted in the selected worktree once a worktree exists.
vbw-fix-planner with the issue body, observed reproduction data, and any user-supplied plan. The planner produces or audits the execution plan before implementation.scripts/ensure-worktree.sh, scripts/adopt-contributor-pr.sh, .github/scripts/fix-issue-record-state.sh, .github/scripts/wait-github.py, and bash testing/run-all.sh.jq for JSON. Do not parse JSON with grep or sed.bash testing/run-all.sh in tail, tee, background execution, or log redirection that hides the real exit status.When finished, report the issue number, branch/worktree, PR URL if created, QA rounds, test command, Codex remote review request status, and CI/review gate status. If blocked, report the exact blocker and the recovery command or next local action.