[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77RDF9DRF477KZJRMS1G1Y5":3,"skill-same:01M77RDF9DRF477KZJRMS1G1Y5":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":29,"updatedAt":30,"install":31,"resources":44},"01M77RDF9DRF477KZJRMS1G1Y5","pr-ship",null,"End-to-end PR lifecycle: create PR, wait for CI, review, fix issues, merge — all in one invocation.\nUse when: (1) User says \"\u002Fpr-ship\", (2) User wants to ship current changes as a complete PR lifecycle,\n(3) User wants to resume shepherding an existing PR.","finewood2008\u002Fcentaurai-team","finewood2008","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F212194470?v=4","01M75HGRZJ31P9W3T90VHFS1X8","Finewood","01M77RC50EAM5Q4E71MX8JRG7P",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002F.claude\u002Fskills\u002Fpr-ship\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: pr-ship\ndescription: |\n  End-to-end PR lifecycle: create PR, wait for CI, review, fix issues, merge — all in one invocation.\n  Use when: (1) User says \"\u002Fpr-ship\", (2) User wants to ship current changes as a complete PR lifecycle,\n  (3) User wants to resume shepherding an existing PR.\n---\n\n# PR Ship\n\nEnd-to-end PR lifecycle shepherd: create PR → wait for CI → review → fix → merge. Single invocation, single PR.\n\n**Announce at start:** \"Using pr-ship skill to shepherd this PR from creation to merge.\"\n\n## Usage\n\n```\n\u002Fpr-ship [pr_number] [--no-auto-merge]\n```\n\n| Parameter         | Default | Description                                                  |\n| ----------------- | ------- | ------------------------------------------------------------ |\n| `pr_number`       | none    | Resume from an existing PR (skip creation)                   |\n| `--no-auto-merge` | off     | Require user confirmation before merge instead of auto-merge |\n\n## Session State\n\nTrack these values in conversation context throughout the session:\n\n- `PR_NUMBER` — current PR number\n- `PHASE` — current phase (create \u002F ci-wait \u002F review \u002F fix \u002F merge)\n- `RETRY_COUNT` — CI failure + review fix retry counter (max 3, shared), initialized to 0\n- `AUTO_MERGE` — true unless `--no-auto-merge` is passed\n- `EMPTY_CI_COUNT` — consecutive empty CI check counter (max 3), initialized to 0\n\n---\n\n## Phase 0 — Create PR\n\n**Skip if `pr_number` is provided.** When skipping, set `PR_NUMBER` from the argument and jump to Phase 1.\n\n**Parse arguments:**\n\n```bash\n# Detect --no-auto-merge flag\nAUTO_MERGE=true\nif echo \"$ARGUMENTS\" | grep -q -- '--no-auto-merge'; then\n  AUTO_MERGE=false\nfi\n\n# Detect pr_number (first numeric argument)\nPR_NUMBER=$(echo \"$ARGUMENTS\" | grep -oE '[0-9]+' | head -1)\n```\n\n**If `PR_NUMBER` is set:** set `RETRY_COUNT = 0`, `EMPTY_CI_COUNT = 0`, skip to Phase 1.\n\n**If `PR_NUMBER` is not set:** invoke oss-pr to create the PR:\n\n```\n\u002Foss-pr\n```\n\nAfter oss-pr completes, extract `PR_NUMBER` from the PR URL in its output. Set `RETRY_COUNT = 0`, `EMPTY_CI_COUNT = 0`.\n\n---\n\n## Phase 1 — CI Wait\n\nCheck CI status:\n\n```bash\ngh pr view $PR_NUMBER --json statusCheckRollup \\\n  --jq '.statusCheckRollup[] | {name: .name, status: .status, conclusion: .conclusion}'\n```\n\n**Required jobs** (same list as pr-review):\n\n- `Code Quality`\n- `Unit Tests (ubuntu-latest)`\n- `Unit Tests (macos-14)`\n- `Unit Tests (windows-2022)`\n- `Coverage Test`\n- `i18n-check`\n\n**Informational exclusions:** `codecov\u002Fpatch` and `codecov\u002Fproject` — always excluded from failure checks.\n\n### Decision Matrix\n\n| CI Status                                                | Action                                            |\n| -------------------------------------------------------- | ------------------------------------------------- |\n| All required jobs SUCCESS, no non-informational failures | → Phase 2                                         |\n| Any required job QUEUED or IN_PROGRESS                   | ScheduleWakeup 270s, re-check                     |\n| `statusCheckRollup` empty (CI never triggered)           | Approve workflow (see below), ScheduleWakeup 270s |\n| Any non-informational job FAILURE or CANCELLED           | → CI Failure Handler                              |\n\n### Workflow Approval (CI not triggered)\n\n```bash\nHEAD_SHA=$(gh pr view $PR_NUMBER --json headRefOid --jq '.headRefOid')\nREPO=$(gh repo view --json nameWithOwner --jq '.nameWithOwner')\nRUN_IDS=$(gh api \"repos\u002F$REPO\u002Factions\u002Fruns?head_sha=$HEAD_SHA&status=action_required\" \\\n  --jq '.workflow_runs[].id')\nfor RUN_ID in $RUN_IDS; do\n  gh run approve \"$RUN_ID\" --repo \"$REPO\"\ndone\n```\n\nThen `EMPTY_CI_COUNT++` and ScheduleWakeup 270s to re-check. If `EMPTY_CI_COUNT >= 3` (~13.5 minutes), abort:\n\n> CI 持续未触发（已等待约 13.5 分钟）。请检查仓库 CI 配置后重新调用 `\u002Fpr-ship $PR_NUMBER`。\n\n### CI Failure Handler\n\n**Step 1 — Check retry budget:**\n\nIf `RETRY_COUNT >= 3`:\n\n> 已达到最大重试次数 (3\u002F3)。请手动检查 PR #$PR_NUMBER 后重新调用 `\u002Fpr-ship $PR_NUMBER`。\n\nExit.\n\n**Step 2 — Identify failures:**\n\n```bash\n# Get failed job names\nFAILED_JOBS=$(gh pr view $PR_NUMBER --json statusCheckRollup \\\n  --jq '[.statusCheckRollup[] | select(.conclusion == \"FAILURE\" or .conclusion == \"CANCELLED\") | select(.name | test(\"^codecov\u002F\") | not) | .name] | join(\", \")')\n```\n\nReport to user:\n\n> CI 失败：$FAILED_JOBS。正在尝试修复 ($RETRY_COUNT\u002F3)...\n\n**Step 3 — Fetch failure details:**\n\n```bash\nREPO=$(gh repo view --json nameWithOwner --jq '.nameWithOwner')\n\n# Find the failed run IDs\nHEAD_SHA=$(gh pr view $PR_NUMBER --json headRefOid --jq '.headRefOid')\nFAILED_RUN_IDS=$(gh api \"repos\u002F$REPO\u002Factions\u002Fruns?head_sha=$HEAD_SHA&status=failure\" \\\n  --jq '.workflow_runs[].id')\n\n# Get logs for each failed run\nfor RUN_ID in $FAILED_RUN_IDS; do\n  gh run view \"$RUN_ID\" --repo \"$REPO\" --log-failed 2>\u002Fdev\u002Fnull | tail -100\ndone\n```\n\n**Step 4 — Fix in worktree:**\n\n```bash\nREPO_ROOT=$(git rev-parse --show-toplevel)\nWORKTREE_DIR=\"\u002Ftmp\u002Faionui-ship-${PR_NUMBER}\"\nHEAD_BRANCH=$(gh pr view $PR_NUMBER --json headRefName --jq '.headRefName')\n\n# Clean up stale worktree\ngit worktree remove \"$WORKTREE_DIR\" --force 2>\u002Fdev\u002Fnull || true\n\n# Create worktree\ngit fetch origin \"$HEAD_BRANCH\"\ngit worktree add \"$WORKTREE_DIR\" \"origin\u002F$HEAD_BRANCH\" --detach\n\n# Symlink node_modules\nln -s \"$REPO_ROOT\u002Fnode_modules\" \"$WORKTREE_DIR\u002Fnode_modules\"\n```\n\nFix only CI-reported errors (lint errors, type errors, test failures) in the worktree. No refactoring, no scope expansion.\n\nAfter fixing, run local quality gate:\n\n```bash\ncd \"$WORKTREE_DIR\"\nbun run lint:fix\nbun run format\nbunx tsc --noEmit\nbun run test\n```\n\n**Step 5 — Commit and push:**\n\n```bash\ncd \"$WORKTREE_DIR\"\ngit add -u\ngit commit -m \"fix(\u003Cscope>): resolve CI failures\"\ngit push origin HEAD:$HEAD_BRANCH\n```\n\n**No AI signature in commits.**\n\n**Step 6 — Cleanup worktree:**\n\n```bash\ncd \"$REPO_ROOT\"\ngit worktree remove \"$WORKTREE_DIR\" --force 2>\u002Fdev\u002Fnull || true\n```\n\n**Step 7 — Increment and loop:**\n\n`RETRY_COUNT++`. ScheduleWakeup 270s → re-enter Phase 1.\n\n---\n\n## Phase 2 — Review\n\nInvoke pr-review in interactive mode, passing the PR number explicitly:\n\n```\n\u002Fpr-review $PR_NUMBER\n```\n\n(No `--automation` flag — this is an interactive session.)\n\n### Decision based on review conclusion\n\n| Review Conclusion                | Action                                            |\n| -------------------------------- | ------------------------------------------------- |\n| ✅ APPROVED (including LOW-only) | → Phase 4                                         |\n| ⚠️ CONDITIONAL                   | → Phase 3                                         |\n| ❌ REJECTED                      | → Phase 3 (pr-fix triage decides if auto-fixable) |\n\n---\n\n## Phase 3 — Fix\n\n**Check retry budget first:**\n\nIf `RETRY_COUNT >= 3`:\n\n> 已达到最大重试次数 (3\u002F3)。请手动检查 PR #$PR_NUMBER 后重新调用 `\u002Fpr-ship $PR_NUMBER`。\n\nExit.\n\nInvoke pr-fix in interactive mode, passing the PR number explicitly:\n\n```\n\u002Fpr-fix $PR_NUMBER\n```\n\n(No `--automation` flag.)\n\n### After pr-fix completes\n\n**Fix pushed successfully:** `RETRY_COUNT++`. ScheduleWakeup 270s → back to Phase 1 (wait for CI on new commit).\n\n**Fix aborted** (all issues dismissed, or user rejected CRITICAL dismissal):\n\n> pr-fix 未产生修复。是否继续看护此 PR？(yes\u002Fno)\n\n- **yes** → back to Phase 1 (CI may already be passing)\n- **no** → exit, output PR URL for manual handling\n\n---\n\n## Phase 4 — Merge\n\n### AUTO_MERGE = true (default)\n\n```bash\ngh pr merge $PR_NUMBER --squash --auto\n```\n\nOutput:\n\n> ✅ PR #$PR_NUMBER 已设置 auto-merge，CI 通过后将自动合并。\n\n### AUTO_MERGE = false (--no-auto-merge)\n\nDisplay summary:\n\n```\n=== PR #$PR_NUMBER 就绪 ===\n\n标题: \u003Ctitle>\n变更: +\u003Cadditions> \u002F -\u003Cdeletions> (\u003Cfile_count> files)\nReview: \u003Cconclusion>\nCI: 全部通过\n\n输入 merge 确认合并，或 abort 放弃。\n```\n\nFetch summary data:\n\n```bash\ngh pr view $PR_NUMBER --json title,additions,deletions,changedFiles \\\n  --jq '{title: .title, additions: .additions, deletions: .deletions, files: .changedFiles}'\n```\n\n- User says **merge** → `gh pr merge $PR_NUMBER --squash` → output `✅ PR #$PR_NUMBER 已合并。`\n- User says **abort** → exit, output PR URL\n\n---\n\n## ScheduleWakeup Integration\n\nThis skill uses `\u002Floop` dynamic mode for CI waiting. Each ScheduleWakeup call:\n\n```\nScheduleWakeup(\n  delaySeconds: 270,\n  reason: \"waiting for CI on PR #$PR_NUMBER (check N)\",\n  prompt: \"\u002Fpr-ship $PR_NUMBER\"\n)\n```\n\n**Interval:** 270s uniformly (under 5-min prompt cache TTL).\n\n**When to call ScheduleWakeup:**\n\n- CI is still running (QUEUED\u002FIN_PROGRESS)\n- Just approved a workflow run\n- Just pushed a CI fix\n- Just pushed a pr-fix commit\n\n**When NOT to call ScheduleWakeup:**\n\n- CI has completed (pass or fail with actionable errors)\n- During review (Phase 2) or fix (Phase 3) execution\n- After merge (Phase 4)\n\n---\n\n## Mandatory Rules\n\n- **No AI signature** — no `Co-Authored-By`, no `Generated with` in any commit or comment\n- **Retry limit = 3** — hard cap, shared between CI fixes and review fixes\n- **Fix scope discipline** — CI fixes target only CI-reported errors; review fixes go through pr-fix\n- **No label pollution** — pr-ship does not set any `bot:*` labels\n- **ScheduleWakeup only for CI wait** — never during review or fix execution\n- **REPO detection at runtime** — `gh repo view --json nameWithOwner --jq '.nameWithOwner'`\n- **Worktree path** — always `\u002Ftmp\u002Faionui-ship-\u003CPR_NUMBER>` (distinct from pr-automation's `\u002Ftmp\u002Faionui-pr-*` and pr-verify's `\u002Ftmp\u002Faionui-verify-*`)\n- **Worktree cleanup** — always remove worktree after CI fix, even on failure\n\n## Quick Reference\n\n```\n0. Parse args: pr_number? --no-auto-merge?\n1. Phase 0: \u002Foss-pr (skip if pr_number given)\n2. Phase 1: CI wait loop\n   - check statusCheckRollup\n   - passing -> Phase 2\n   - running -> ScheduleWakeup 270s\n   - not triggered -> approve workflow + ScheduleWakeup 270s\n   - failed -> retry_count \u003C 3? fix + push + ScheduleWakeup 270s : abort\n3. Phase 2: \u002Fpr-review (interactive)\n   - APPROVED -> Phase 4\n   - CONDITIONAL\u002FREJECTED -> Phase 3\n4. Phase 3: \u002Fpr-fix (interactive)\n   - pushed -> retry_count++, back to Phase 1\n   - aborted -> ask user\n5. Phase 4: merge\n   - default: gh pr merge --squash --auto\n   - --no-auto-merge: confirm then gh pr merge --squash\n```\n","# PR Ship\n\nEnd-to-end PR lifecycle shepherd: create PR → wait for CI → review → fix → merge. Single invocation, single PR.\n\n**Announce at start:** \"Using pr-ship skill to shepherd this PR from creation to merge.\"\n\n## Usage\n\n```\n\u002Fpr-ship [pr_number] [--no-auto-merge]\n```\n\n| Parameter         | Default | Description                                                  |\n| ----------------- | ------- | ------------------------------------------------------------ |\n| `pr_number`       | none    | Resume from an existing PR (skip creation)                   |\n| `--no-auto-merge` | off     | Require user confirmation before merge instead of auto-merge |\n\n## Session State\n\nTrack these values in conversation context throughout the session:\n\n- `PR_NUMBER` — current PR number\n- `PHASE` — current phase (create \u002F ci-wait \u002F review \u002F fix \u002F merge)\n- `RETRY_COUNT` — CI failure + review fix retry counter (max 3, shared), initialized to 0\n- `AUTO_MERGE` — true unless `--no-auto-merge` is passed\n- `EMPTY_CI_COUNT` — consecutive empty CI check counter (max 3), initialized to 0\n\n---\n\n## Phase 0 — Create PR\n\n**Skip if `pr_number` is provided.** When skipping, set `PR_NUMBER` from the argument and jump to Phase 1.\n\n**Parse arguments:**\n\n```bash\n# Detect --no-auto-merge flag\nAUTO_MERGE=true\nif echo \"$ARGUMENTS\" | grep -q -- '--no-auto-merge'; then\n  AUTO_MERGE=false\nfi\n\n# Detect pr_number (first numeric argument)\nPR_NUMBER=$(echo \"$ARGUMENTS\" | grep -oE '[0-9]+' | head -1)\n```\n\n**If `PR_NUMBER` is set:** set `RETRY_COUNT = 0`, `EMPTY_CI_COUNT = 0`, skip to Phase 1.\n\n**If `PR_NUMBER` is not set:** invoke oss-pr to create the PR:\n\n```\n\u002Foss-pr\n```\n\nAfter oss-pr completes, extract `PR_NUMBER` from the PR URL in its output. Set `RETRY_COUNT = 0`, `EMPTY_CI_COUNT = 0`.\n\n---\n\n## Phase 1 — CI Wait\n\nCheck CI status:\n\n```bash\ngh pr view $PR_NUMBER --json statusCheckRollup \\\n  --jq '.statusCheckRollup[] | {name: .name, status: .status, conclusion: .conclusion}'\n```\n\n**Required jobs** (same list as pr-review):\n\n- `Code Quality`\n- `Unit Tests (ubuntu-latest)`\n- `Unit Tests (macos-14)`\n- `Unit Tests (windows-2022)`\n- `Coverage Test`\n- `i18n-check`\n\n**Informational exclusions:** `codecov\u002Fpatch` and `codecov\u002Fproject` — always excluded from failure checks.\n\n### Decision Matrix\n\n| CI Status                                                | Action                                            |\n| -------------------------------------------------------- | ------------------------------------------------- |\n| All required jobs SUCCESS, no non-informational failures | → Phase 2                                         |\n| Any required job QUEUED or IN_PROGRESS                   | ScheduleWakeup 270s, re-check                     |\n| `statusCheckRollup` empty (CI never triggered)           | Approve workflow (see below), ScheduleWakeup 270s |\n| Any non-informational job FAILURE or CANCELLED           | → CI Failure Handler                              |\n\n### Workflow Approval (CI not triggered)\n\n```bash\nHEAD_SHA=$(gh pr view $PR_NUMBER --json headRefOid --jq '.headRefOid')\nREPO=$(gh repo view --json nameWithOwner --jq '.nameWithOwner')\nRUN_IDS=$(gh api \"repos\u002F$REPO\u002Factions\u002Fruns?head_sha=$HEAD_SHA&status=action_required\" \\\n  --jq '.workflow_runs[].id')\nfor RUN_ID in $RUN_IDS; do\n  gh run approve \"$RUN_ID\" --repo \"$REPO\"\ndone\n```\n\nThen `EMPTY_CI_COUNT++` and ScheduleWakeup 270s to re-check. If `EMPTY_CI_COUNT >= 3` (~13.5 minutes), abort:\n\n> CI 持续未触发（已等待约 13.5 分钟）。请检查仓库 CI 配置后重新调用 `\u002Fpr-ship $PR_NUMBER`。\n\n### CI Failure Handler\n\n**Step 1 — Check retry budget:**\n\nIf `RETRY_COUNT >= 3`:\n\n> 已达到最大重试次数 (3\u002F3)。请手动检查 PR #$PR_NUMBER 后重新调用 `\u002Fpr-ship $PR_NUMBER`。\n\nExit.\n\n**Step 2 — Identify failures:**\n\n```bash\n# Get failed job names\nFAILED_JOBS=$(gh pr view $PR_NUMBER --json statusCheckRollup \\\n  --jq '[.statusCheckRollup[] | select(.conclusion == \"FAILURE\" or .conclusion == \"CANCELLED\") | select(.name | test(\"^codecov\u002F\") | not) | .name] | join(\", \")')\n```\n\nReport to user:\n\n> CI 失败：$FAILED_JOBS。正在尝试修复 ($RETRY_COUNT\u002F3)...\n\n**Step 3 — Fetch failure details:**\n\n```bash\nREPO=$(gh repo view --json nameWithOwner --jq '.nameWithOwner')\n\n# Find the failed run IDs\nHEAD_SHA=$(gh pr view $PR_NUMBER --json headRefOid --jq '.headRefOid')\nFAILED_RUN_IDS=$(gh api \"repos\u002F$REPO\u002Factions\u002Fruns?head_sha=$HEAD_SHA&status=failure\" \\\n  --jq '.workflow_runs[].id')\n\n# Get logs for each failed run\nfor RUN_ID in $FAILED_RUN_IDS; do\n  gh run view \"$RUN_ID\" --repo \"$REPO\" --log-failed 2>\u002Fdev\u002Fnull | tail -100\ndone\n```\n\n**Step 4 — Fix in worktree:**\n\n```bash\nREPO_ROOT=$(git rev-parse --show-toplevel)\nWORKTREE_DIR=\"\u002Ftmp\u002Faionui-ship-${PR_NUMBER}\"\nHEAD_BRANCH=$(gh pr view $PR_NUMBER --json headRefName --jq '.headRefName')\n\n# Clean up stale worktree\ngit worktree remove \"$WORKTREE_DIR\" --force 2>\u002Fdev\u002Fnull || true\n\n# Create worktree\ngit fetch origin \"$HEAD_BRANCH\"\ngit worktree add \"$WORKTREE_DIR\" \"origin\u002F$HEAD_BRANCH\" --detach\n\n# Symlink node_modules\nln -s \"$REPO_ROOT\u002Fnode_modules\" \"$WORKTREE_DIR\u002Fnode_modules\"\n```\n\nFix only CI-reported errors (lint errors, type errors, test failures) in the worktree. No refactoring, no scope expansion.\n\nAfter fixing, run local quality gate:\n\n```bash\ncd \"$WORKTREE_DIR\"\nbun run lint:fix\nbun run format\nbunx tsc --noEmit\nbun run test\n```\n\n**Step 5 — Commit and push:**\n\n```bash\ncd \"$WORKTREE_DIR\"\ngit add -u\ngit commit -m \"fix(\u003Cscope>): resolve CI failures\"\ngit push origin HEAD:$HEAD_BRANCH\n```\n\n**No AI signature in commits.**\n\n**Step 6 — Cleanup worktree:**\n\n```bash\ncd \"$REPO_ROOT\"\ngit worktree remove \"$WORKTREE_DIR\" --force 2>\u002Fdev\u002Fnull || true\n```\n\n**Step 7 — Increment and loop:**\n\n`RETRY_COUNT++`. ScheduleWakeup 270s → re-enter Phase 1.\n\n---\n\n## Phase 2 — Review\n\nInvoke pr-review in interactive mode, passing the PR number explicitly:\n\n```\n\u002Fpr-review $PR_NUMBER\n```\n\n(No `--automation` flag — this is an interactive session.)\n\n### Decision based on review conclusion\n\n| Review Conclusion                | Action                                            |\n| -------------------------------- | ------------------------------------------------- |\n| ✅ APPROVED (including LOW-only) | → Phase 4                                         |\n| ⚠️ CONDITIONAL                   | → Phase 3                                         |\n| ❌ REJECTED                      | → Phase 3 (pr-fix triage decides if auto-fixable) |\n\n---\n\n## Phase 3 — Fix\n\n**Check retry budget first:**\n\nIf `RETRY_COUNT >= 3`:\n\n> 已达到最大重试次数 (3\u002F3)。请手动检查 PR #$PR_NUMBER 后重新调用 `\u002Fpr-ship $PR_NUMBER`。\n\nExit.\n\nInvoke pr-fix in interactive mode, passing the PR number explicitly:\n\n```\n\u002Fpr-fix $PR_NUMBER\n```\n\n(No `--automation` flag.)\n\n### After pr-fix completes\n\n**Fix pushed successfully:** `RETRY_COUNT++`. ScheduleWakeup 270s → back to Phase 1 (wait for CI on new commit).\n\n**Fix aborted** (all issues dismissed, or user rejected CRITICAL dismissal):\n\n> pr-fix 未产生修复。是否继续看护此 PR？(yes\u002Fno)\n\n- **yes** → back to Phase 1 (CI may already be passing)\n- **no** → exit, output PR URL for manual handling\n\n---\n\n## Phase 4 — Merge\n\n### AUTO_MERGE = true (default)\n\n```bash\ngh pr merge $PR_NUMBER --squash --auto\n```\n\nOutput:\n\n> ✅ PR #$PR_NUMBER 已设置 auto-merge，CI 通过后将自动合并。\n\n### AUTO_MERGE = false (--no-auto-merge)\n\nDisplay summary:\n\n```\n=== PR #$PR_NUMBER 就绪 ===\n\n标题: \u003Ctitle>\n变更: +\u003Cadditions> \u002F -\u003Cdeletions> (\u003Cfile_count> files)\nReview: \u003Cconclusion>\nCI: 全部通过\n\n输入 merge 确认合并，或 abort 放弃。\n```\n\nFetch summary data:\n\n```bash\ngh pr view $PR_NUMBER --json title,additions,deletions,changedFiles \\\n  --jq '{title: .title, additions: .additions, deletions: .deletions, files: .changedFiles}'\n```\n\n- User says **merge** → `gh pr merge $PR_NUMBER --squash` → output `✅ PR #$PR_NUMBER 已合并。`\n- User says **abort** → exit, output PR URL\n\n---\n\n## ScheduleWakeup Integration\n\nThis skill uses `\u002Floop` dynamic mode for CI waiting. Each ScheduleWakeup call:\n\n```\nScheduleWakeup(\n  delaySeconds: 270,\n  reason: \"waiting for CI on PR #$PR_NUMBER (check N)\",\n  prompt: \"\u002Fpr-ship $PR_NUMBER\"\n)\n```\n\n**Interval:** 270s uniformly (under 5-min prompt cache TTL).\n\n**When to call ScheduleWakeup:**\n\n- CI is still running (QUEUED\u002FIN_PROGRESS)\n- Just approved a workflow run\n- Just pushed a CI fix\n- Just pushed a pr-fix commit\n\n**When NOT to call ScheduleWakeup:**\n\n- CI has completed (pass or fail with actionable errors)\n- During review (Phase 2) or fix (Phase 3) execution\n- After merge (Phase 4)\n\n---\n\n## Mandatory Rules\n\n- **No AI signature** — no `Co-Authored-By`, no `Generated with` in any commit or comment\n- **Retry limit = 3** — hard cap, shared between CI fixes and review fixes\n- **Fix scope discipline** — CI fixes target only CI-reported errors; review fixes go through pr-fix\n- **No label pollution** — pr-ship does not set any `bot:*` labels\n- **ScheduleWakeup only for CI wait** — never during review or fix execution\n- **REPO detection at runtime** — `gh repo view --json nameWithOwner --jq '.nameWithOwner'`\n- **Worktree path** — always `\u002Ftmp\u002Faionui-ship-\u003CPR_NUMBER>` (distinct from pr-automation's `\u002Ftmp\u002Faionui-pr-*` and pr-verify's `\u002Ftmp\u002Faionui-verify-*`)\n- **Worktree cleanup** — always remove worktree after CI fix, even on failure\n\n## Quick Reference\n\n```\n0. Parse args: pr_number? --no-auto-merge?\n1. Phase 0: \u002Foss-pr (skip if pr_number given)\n2. Phase 1: CI wait loop\n   - check statusCheckRollup\n   - passing -> Phase 2\n   - running -> ScheduleWakeup 270s\n   - not triggered -> approve workflow + ScheduleWakeup 270s\n   - failed -> retry_count \u003C 3? fix + push + ScheduleWakeup 270s : abort\n3. Phase 2: \u002Fpr-review (interactive)\n   - APPROVED -> Phase 4\n   - CONDITIONAL\u002FREJECTED -> Phase 3\n4. Phase 3: \u002Fpr-fix (interactive)\n   - pushed -> retry_count++, back to Phase 1\n   - aborted -> ask user\n5. Phase 4: merge\n   - default: gh pr merge --squash --auto\n   - --no-auto-merge: confirm then gh pr merge --squash\n```\n",{"name":5,"description":7},"team","a9a9464c62247a811d9231c636a0941e0fd9ae05",".claude\u002Fskills\u002Fpr-ship","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team","https:\u002F\u002Fgithub.com\u002Ffinewood2008","Founder & CEO of Centaur AI · Humans lead, AI executes — the Centaur Loop is the smallest work unit for building AI employees.","2026-07-14 22:41:28",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add finewood2008\u002Fcentaurai-team --skill pr-ship",{"tool":37,"command":38},"bunx","bunx skills add finewood2008\u002Fcentaurai-team --skill pr-ship",{"tool":40,"command":41},"pnpm","pnpm dlx skills add finewood2008\u002Fcentaurai-team --skill pr-ship","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002F.claude\u002Fskills\u002Fpr-ship\u002FSKILL.md",true,[],[46,54,61,68,75,82],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M77RFHB8MZ42KXGPY38CXCAQ","团队周报","Use when generating structured weekly team reports from git commits, task boards, and meeting notes. Auto-aggregates work done, work planned, blockers, and metrics. Chinese output by default. Trigger on: weekly report, 周报, sprint review, 本周总结, team update, status report.","MIT","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002Fskills\u002Fbuiltin\u002Fteam\u002Fteam-weekly-report\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":19},"01M77RFBY70Q5R4YN8A3JD7X5G","中英互译","Use when translating content between Chinese and English for business, technical, or marketing purposes. Handles docs, emails, UI strings, marketing copy, and technical documentation with domain-aware terminology. Trigger on: 翻译, translate, 中译英, 英译中, i18n, localization, English version, 中文版.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002Fskills\u002Fbuiltin\u002Fteam\u002Fteam-translation\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M77RF6YBVG39A4QQQEGY8Q6N","社媒规划","Use when planning, drafting, or scheduling social media content for WeChat Official Account, Twitter\u002FX, and Xiaohongshu. Generates content calendars with platform-adapted copy, publishing schedules, and performance tracking. Trigger on: social media, 公众号, 推文, content calendar, 社媒, 发布计划.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002Fskills\u002Fbuiltin\u002Fteam\u002Fteam-social-planner\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":19},"01M77RF2JR9BZ9QKXY101Z80F8","项目任务看板","Use when managing team projects, tasks, or sprint tracking with a lightweight markdown kanban board. Create, update, assign, and track tasks across team members without external tools. Trigger on: task board, kanban, sprint planning, assign task, task status, project progress.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002Fskills\u002Fbuiltin\u002Fteam\u002Fteam-project-tracker\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M77REYK38P4W54XQVFFV0GF2","新人入职","Use when onboarding a new team member to the company, project, or development environment. Covers environment setup, account provisioning, access permissions, first-day tasks, and knowledge base orientation. Trigger on: onboarding, 入职, new employee, 新人, setup workstation, 环境配置, first day.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002Fskills\u002Fbuiltin\u002Fteam\u002Fteam-onboarding\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":19},"01M77RESXY98BF66BSTPZADB4W","会议纪要","Use when recording, structuring, or extracting action items from team meetings. Creates standardized meeting minutes with decisions, action items with owners and deadlines, and follow-up tracking. Trigger on: meeting notes, 会议纪要, meeting minutes, 开会记录, standup notes, decision log, 会议记录.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-team\u002Fblob\u002Fteam\u002Fskills\u002Fbuiltin\u002Fteam\u002Fteam-meeting-minutes\u002FSKILL.md",[],[]]