[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6C31K6MSYN405JPXVF4VZRW":3,"skill-same:01M6C31K6MSYN405JPXVF4VZRW":75},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":22,"publishedAt":48,"linkOnly":49,"markdownContent":50,"markdownBody":51,"frontMatter":52,"refName":54,"commitSha":55,"skillDirectory":56,"forksCount":57,"repositoryHtmlUrl":58,"repositoryHomepageUrl":6,"ownerHtmlUrl":59,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":60,"install":61,"resources":74},"01M6C31K6MSYN405JPXVF4VZRW","ao-worker-ground-truth",null,"Diagnose \"is AO actually working?\" by cross-referencing live OS processes,\nthe AO database (`~\u002F.ao\u002Fdata\u002Fao.db`), and GitHub PR commit history — bypassing\nthe `ao-progress-reporter.sh` and MCP Agent Mail daily report, both of which\ncan show stale or fabricated session data. Use when user asks \"is AO working?\",\n\"are AO workers finishing?\", or \"the daily report shows N stalled\".","jleechanorg\u002Fjleechan-skills","jleechanorg","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F221404966?v=4","01M6C1VJ99XG950X1DWPEBCBGX","01M6C1VJDK5YVDNT3362RT5WSB",6,"MIT","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002Fhermes\u002Fskills\u002Fao-worker-ground-truth\u002FSKILL.md",[17],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47],"worker","ao","ao worker ground truth","daily report","when","which","progress","os","db","live","mail","show","mcp agent mail","MCP","use","session","user","history","data","database","report","pr","diagnose","commit","github","2026-11-10 10:12:30",false,"---\nname: ao-worker-ground-truth\nversion: 1.0.0\ndescription: |\n  Diagnose \"is AO actually working?\" by cross-referencing live OS processes,\n  the AO database (`~\u002F.ao\u002Fdata\u002Fao.db`), and GitHub PR commit history — bypassing\n  the `ao-progress-reporter.sh` and MCP Agent Mail daily report, both of which\n  can show stale or fabricated session data. Use when user asks \"is AO working?\",\n  \"are AO workers finishing?\", or \"the daily report shows N stalled\".\n---\n\n# AO Worker Ground Truth — bypassing stale reports to verify real progress\n\n## Why this skill exists\n\nThere are **three independent sources** for \"what AO workers are doing,\" and\neach can lie in a different way:\n\n| Source | Can be wrong because |\n|---|---|\n| `~\u002F.hermes\u002Fscripts\u002Fao-progress-reporter.sh` daily Slack report | Only classifies sessions with `branch` + `repo` recorded. Sessions with `branch=\"\"` fall through invisibly and aren't reported. |\n| MCP Agent Mail `U0A4G7LDJ4R` \"AO Progress Report\" thread | Uses fabricated `wa-NNNN` IDs that don't match any row in `~\u002F.ao\u002Fdata\u002Fao.db`. Verified 2026-07-23 — bead `orch-1oli`. |\n| `ao session ls` (canonical AO DB) | Has real session IDs but lacks `branch`\u002F`prNumber` for most sessions. |\n\nWhen the user asks \"is AO working?\" or \"is PR #N actually being driven?\", the\nright move is **don't trust any of those three alone**. Instead, triangulate\nfrom a 4th source: **the actual OS processes running on the machine**, plus\nthe **GitHub PR commit history** for the PRs those processes reference.\n\n## The 4-step ground-truth recipe\n\n### Step 1 — Count live AO workers by project\n\n```bash\nao session ls --json | python3 -c \"\nimport json, sys\nd = json.load(sys.stdin)\nsessions = d.get('data', [])\nlive = [s for s in sessions if not s.get('isTerminated')]\nprint(f'Live AO workers: {len(live)}')\nby_proj = {}\nfor s in live:\n    p = s.get('projectId', '?')\n    by_proj.setdefault(p, []).append(s['id'])\nfor p, ids in sorted(by_proj.items()):\n    print(f'  {p}: {len(ids)} ({ids[:5]}{\\\"...\\\" if len(ids) > 5 else \\\"\\\"})')\n\"\n```\n\n### Step 2 — Map live OS processes to PR\u002Fbead IDs (THE KEY STEP)\n\n```bash\nps aux | grep -E \"(claude|agy|codex).*--prompt-interactive\" | grep -v grep \\\n  | grep -oE 'PR #[0-9]+|$USER-[a-z0-9]+|feat\u002F[a-zA-Z0-9_-]+|fix\u002F[a-zA-Z0-9_-]+' \\\n  | sort | uniq -c | sort -rn\n```\n\nThis is the diagnostic that bypasses the AO DB's `branch=\"\"` gap entirely.\nEvery long-running AO worker has a `--prompt-interactive` invocation whose\nembedded brief names the target PR + bead. Pulling the strings out of `ps aux`\ngives you the **actual current work assignments** without depending on AO's\nsession-tracking state.\n\nIf the user asked \"is PR #8178 being driven?\" and this command shows a\n`PR #8178` token in some process's args, the answer is **yes, right now**.\n\n### Step 3 — Cross-reference with GitHub PR commit history\n\nFor each PR the user cares about:\n\n```bash\nfor pr in 8178 8177 8511 8428; do\n  echo \"=== PR #$pr ===\"\n  gh api \"repos\u002Fjleechanorg\u002F\u003Crepo>\u002Fpulls\u002F$pr\u002Fcommits\" | python3 -c \"\nimport json, sys, datetime\nd = json.load(sys.stdin)\nnow = datetime.datetime.now(datetime.timezone.utc)\nfor c in d[-5:]:\n    sha = c.get('sha', '')[:10]\n    date = c.get('commit', {}).get('author', {}).get('date', '')\n    msg = (c.get('commit', {}).get('message', '') or '').split('\\n')[0][:80]\n    if date:\n        dt = datetime.datetime.fromisoformat(date.replace('Z', '+00:00'))\n        age_h = (now - dt).total_seconds() \u002F 3600\n        print(f'  {sha} {age_h:7.1f}h ago  {msg}')\n\"\ndone\n```\n\nThis is the second half of the triangulation: even if `ps aux` shows nothing\n(the worker was killed, or is between spawn and first prompt), a fresh commit\non `origin\u002F\u003Cbranch>` proves the worker was active minutes ago.\n\n**Key sanity check:** if the daily report says \"PR #8511 idle 49h21m\" but this\ncommand shows the head commit is 1.0h old, the daily report is wrong. Do not\ntrust the report — trust GitHub.\n\n### Step 4 — Cross-check the AO DB's `pr` table for canonical state\n\n```bash\nsqlite3 ~\u002F.ao\u002Fdata\u002Fao.db \\\n  \"SELECT number, pr_state, ci_state, mergeability, source_branch, updated_at\n   FROM pr WHERE number IN (8178, 8177, 8511, 8428) ORDER BY updated_at DESC;\"\n```\n\nIf the PR is in the `pr` table, AO has attached itself to it (`ao claim-pr`\nwas called at some point). If the PR is NOT in the table but appears in\n`ps aux` Step 2 output, the worker is mid-flight on a fresh PR without a prior\n`claim-pr`. That's normal and expected.\n\n## Worked example — 2026-07-23 incident\n\n**User asked in thread C0ALSKLU9KM\u002Fp1784792447.282019:**\n\n> \"Is AO actually working? Look at slack history and see if these AO\n> dispatches ever finished their work.\"\n\n**MCP Mail report said:**\n- `wa-3359`: PR open @ `72b4a1a`, off-track, idle 65h48m\n- `wa-3361`: PR open @ `aa92321`, Green Gate FAILURE, idle 63h3m\n- `wa-3364`: PR #8511, off-track, idle 49h21m\n- `wa-3365`: PR #8428, on-track, beads-jsonl-validation IN_PROGRESS\n\n**Ground truth recipe output:**\n\n| Source | Result |\n|---|---|\n| Step 1 (`ao session ls`) | 32 live workers; `worldarchitect-79`, `-88`, `-89` all `no_signal` with `branch=\"\"` |\n| Step 2 (`ps aux`) | 15+ processes with `--prompt-interactive` containing `PR #8178`, `PR #8177`, `PR #8428`, `PR #8466`, `PR #8477`, `PR #8527`, `PR #8529`, `PR #8530`, `PR #63`, `PR #8116` |\n| Step 3 (GitHub) | PR #8511 last commit **1.0h ago**; PR #8428 last 3 commits **1.5–1.6h ago** |\n| Step 4 (`pr` table) | PRs #8178, #8177, #8511, #8428 are **NOT in the `pr` table** — AO never `claim-pr`'d them |\n\n**Verdict:** AO IS working. 15+ workers are actively pushing commits to the\nPRs the daily report called \"stalled.\" The daily report is wrong because:\n1. MCP Mail uses fabricated `wa-NNNN` IDs that don't map to real sessions.\n2. The local reporter falls through on `branch=\"\"` workers.\n3. The AO `pr` table is empty for in-flight PRs (AO never claimed them).\n\n## Anti-patterns\n\n- **Do NOT trust `wa-NNNN` IDs from MCP Mail.** Verify against `ao session ls`\n  and the SQLite `sessions.issue_id` column first. If the ID isn't in either,\n  it's fabricated.\n- **Do NOT trust \"idle N hours\" claims without checking GitHub.** Green Gate\n  workflow idle time is NOT worker idle time — a PR at head SHA `X` for 7 days\n  is **waiting for CI**, not stalled.\n- **Do NOT spawn a new worker when the diagnostic shows the existing one is\n  alive and pushing.** \"Idle in MCP report\" + \"fresh commits in git log\"\n  means the report is wrong, not the worker. Fix the report; don't dispatch.\n- **Do NOT depend on the AO `pr` table for in-flight PRs.** AO workers\n  routinely push to PRs without `claim-pr` first; the `pr` table only\n  reflects explicit claims, which is a strict subset of real work.\n\n## Related skills\n\n- `agento` — how to dispatch + monitor AO workers (canonical CLI surface).\n- `agento_report` — getting a PR-status snapshot; doesn't validate against\n  live processes (use this skill instead when the user says \"really\" or\n  \"actually\").\n- `mcp-agent-mail-no-slack-bridge\u002Freferences\u002Ftwo-identity-slack-routing.md` —\n  documents the `wa-NNNN` fabrication specifically; this skill documents the\n  ground-truth recipe that bypasses it.\n- `slack-thread-routing-investigation` — Failure 5 (wrong thread_ts from\n  session context header) compounds with this: the same header that misroutes\n  your reply can also be the source of the `wa-NNNN` data you're verifying.\n","# AO Worker Ground Truth — bypassing stale reports to verify real progress\n\n## Why this skill exists\n\nThere are **three independent sources** for \"what AO workers are doing,\" and\neach can lie in a different way:\n\n| Source | Can be wrong because |\n|---|---|\n| `~\u002F.hermes\u002Fscripts\u002Fao-progress-reporter.sh` daily Slack report | Only classifies sessions with `branch` + `repo` recorded. Sessions with `branch=\"\"` fall through invisibly and aren't reported. |\n| MCP Agent Mail `U0A4G7LDJ4R` \"AO Progress Report\" thread | Uses fabricated `wa-NNNN` IDs that don't match any row in `~\u002F.ao\u002Fdata\u002Fao.db`. Verified 2026-07-23 — bead `orch-1oli`. |\n| `ao session ls` (canonical AO DB) | Has real session IDs but lacks `branch`\u002F`prNumber` for most sessions. |\n\nWhen the user asks \"is AO working?\" or \"is PR #N actually being driven?\", the\nright move is **don't trust any of those three alone**. Instead, triangulate\nfrom a 4th source: **the actual OS processes running on the machine**, plus\nthe **GitHub PR commit history** for the PRs those processes reference.\n\n## The 4-step ground-truth recipe\n\n### Step 1 — Count live AO workers by project\n\n```bash\nao session ls --json | python3 -c \"\nimport json, sys\nd = json.load(sys.stdin)\nsessions = d.get('data', [])\nlive = [s for s in sessions if not s.get('isTerminated')]\nprint(f'Live AO workers: {len(live)}')\nby_proj = {}\nfor s in live:\n    p = s.get('projectId', '?')\n    by_proj.setdefault(p, []).append(s['id'])\nfor p, ids in sorted(by_proj.items()):\n    print(f'  {p}: {len(ids)} ({ids[:5]}{\\\"...\\\" if len(ids) > 5 else \\\"\\\"})')\n\"\n```\n\n### Step 2 — Map live OS processes to PR\u002Fbead IDs (THE KEY STEP)\n\n```bash\nps aux | grep -E \"(claude|agy|codex).*--prompt-interactive\" | grep -v grep \\\n  | grep -oE 'PR #[0-9]+|$USER-[a-z0-9]+|feat\u002F[a-zA-Z0-9_-]+|fix\u002F[a-zA-Z0-9_-]+' \\\n  | sort | uniq -c | sort -rn\n```\n\nThis is the diagnostic that bypasses the AO DB's `branch=\"\"` gap entirely.\nEvery long-running AO worker has a `--prompt-interactive` invocation whose\nembedded brief names the target PR + bead. Pulling the strings out of `ps aux`\ngives you the **actual current work assignments** without depending on AO's\nsession-tracking state.\n\nIf the user asked \"is PR #8178 being driven?\" and this command shows a\n`PR #8178` token in some process's args, the answer is **yes, right now**.\n\n### Step 3 — Cross-reference with GitHub PR commit history\n\nFor each PR the user cares about:\n\n```bash\nfor pr in 8178 8177 8511 8428; do\n  echo \"=== PR #$pr ===\"\n  gh api \"repos\u002Fjleechanorg\u002F\u003Crepo>\u002Fpulls\u002F$pr\u002Fcommits\" | python3 -c \"\nimport json, sys, datetime\nd = json.load(sys.stdin)\nnow = datetime.datetime.now(datetime.timezone.utc)\nfor c in d[-5:]:\n    sha = c.get('sha', '')[:10]\n    date = c.get('commit', {}).get('author', {}).get('date', '')\n    msg = (c.get('commit', {}).get('message', '') or '').split('\\n')[0][:80]\n    if date:\n        dt = datetime.datetime.fromisoformat(date.replace('Z', '+00:00'))\n        age_h = (now - dt).total_seconds() \u002F 3600\n        print(f'  {sha} {age_h:7.1f}h ago  {msg}')\n\"\ndone\n```\n\nThis is the second half of the triangulation: even if `ps aux` shows nothing\n(the worker was killed, or is between spawn and first prompt), a fresh commit\non `origin\u002F\u003Cbranch>` proves the worker was active minutes ago.\n\n**Key sanity check:** if the daily report says \"PR #8511 idle 49h21m\" but this\ncommand shows the head commit is 1.0h old, the daily report is wrong. Do not\ntrust the report — trust GitHub.\n\n### Step 4 — Cross-check the AO DB's `pr` table for canonical state\n\n```bash\nsqlite3 ~\u002F.ao\u002Fdata\u002Fao.db \\\n  \"SELECT number, pr_state, ci_state, mergeability, source_branch, updated_at\n   FROM pr WHERE number IN (8178, 8177, 8511, 8428) ORDER BY updated_at DESC;\"\n```\n\nIf the PR is in the `pr` table, AO has attached itself to it (`ao claim-pr`\nwas called at some point). If the PR is NOT in the table but appears in\n`ps aux` Step 2 output, the worker is mid-flight on a fresh PR without a prior\n`claim-pr`. That's normal and expected.\n\n## Worked example — 2026-07-23 incident\n\n**User asked in thread C0ALSKLU9KM\u002Fp1784792447.282019:**\n\n> \"Is AO actually working? Look at slack history and see if these AO\n> dispatches ever finished their work.\"\n\n**MCP Mail report said:**\n- `wa-3359`: PR open @ `72b4a1a`, off-track, idle 65h48m\n- `wa-3361`: PR open @ `aa92321`, Green Gate FAILURE, idle 63h3m\n- `wa-3364`: PR #8511, off-track, idle 49h21m\n- `wa-3365`: PR #8428, on-track, beads-jsonl-validation IN_PROGRESS\n\n**Ground truth recipe output:**\n\n| Source | Result |\n|---|---|\n| Step 1 (`ao session ls`) | 32 live workers; `worldarchitect-79`, `-88`, `-89` all `no_signal` with `branch=\"\"` |\n| Step 2 (`ps aux`) | 15+ processes with `--prompt-interactive` containing `PR #8178`, `PR #8177`, `PR #8428`, `PR #8466`, `PR #8477`, `PR #8527`, `PR #8529`, `PR #8530`, `PR #63`, `PR #8116` |\n| Step 3 (GitHub) | PR #8511 last commit **1.0h ago**; PR #8428 last 3 commits **1.5–1.6h ago** |\n| Step 4 (`pr` table) | PRs #8178, #8177, #8511, #8428 are **NOT in the `pr` table** — AO never `claim-pr`'d them |\n\n**Verdict:** AO IS working. 15+ workers are actively pushing commits to the\nPRs the daily report called \"stalled.\" The daily report is wrong because:\n1. MCP Mail uses fabricated `wa-NNNN` IDs that don't map to real sessions.\n2. The local reporter falls through on `branch=\"\"` workers.\n3. The AO `pr` table is empty for in-flight PRs (AO never claimed them).\n\n## Anti-patterns\n\n- **Do NOT trust `wa-NNNN` IDs from MCP Mail.** Verify against `ao session ls`\n  and the SQLite `sessions.issue_id` column first. If the ID isn't in either,\n  it's fabricated.\n- **Do NOT trust \"idle N hours\" claims without checking GitHub.** Green Gate\n  workflow idle time is NOT worker idle time — a PR at head SHA `X` for 7 days\n  is **waiting for CI**, not stalled.\n- **Do NOT spawn a new worker when the diagnostic shows the existing one is\n  alive and pushing.** \"Idle in MCP report\" + \"fresh commits in git log\"\n  means the report is wrong, not the worker. Fix the report; don't dispatch.\n- **Do NOT depend on the AO `pr` table for in-flight PRs.** AO workers\n  routinely push to PRs without `claim-pr` first; the `pr` table only\n  reflects explicit claims, which is a strict subset of real work.\n\n## Related skills\n\n- `agento` — how to dispatch + monitor AO workers (canonical CLI surface).\n- `agento_report` — getting a PR-status snapshot; doesn't validate against\n  live processes (use this skill instead when the user says \"really\" or\n  \"actually\").\n- `mcp-agent-mail-no-slack-bridge\u002Freferences\u002Ftwo-identity-slack-routing.md` —\n  documents the `wa-NNNN` fabrication specifically; this skill documents the\n  ground-truth recipe that bypasses it.\n- `slack-thread-routing-investigation` — Failure 5 (wrong thread_ts from\n  session context header) compounds with this: the same header that misroutes\n  your reply can also be the source of the `wa-NNNN` data you're verifying.\n",{"name":5,"version":53,"description":7},"1.0.0","main","42ccd02cbf5e8e47839dc3eab6fc754dc6b339cf","hermes\u002Fskills\u002Fao-worker-ground-truth",0,"https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills","https:\u002F\u002Fgithub.com\u002Fjleechanorg","2026-11-14 03:18:13",{"source":8,"skillName":5,"commands":62,"prompt":72,"zipAvailable":73},[63,66,69],{"tool":64,"command":65},"npx","npx skills add jleechanorg\u002Fjleechan-skills --skill ao-worker-ground-truth",{"tool":67,"command":68},"bunx","bunx skills add jleechanorg\u002Fjleechan-skills --skill ao-worker-ground-truth",{"tool":70,"command":71},"pnpm","pnpm dlx skills add jleechanorg\u002Fjleechan-skills --skill ao-worker-ground-truth","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002Fhermes\u002Fskills\u002Fao-worker-ground-truth\u002FSKILL.md",true,[],[76,91,112,126,140,156],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":48},"01M716Y5MTJ6GYJZ2ATXCJK8G4","instinct","Use when the user invokes \u002Finstinct or \u002Fin, or asks to \"message Instinct\", \"ask Instinct\", \"tell Instinct\", \"check Instinct\", or read replies from the configured peer assistant. iMessage first, WhatsApp Web as fallback.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Finstinct\u002FSKILL.md",[],[78,39,27,83,84,37,85,86,87,88,89,90],"replies","check","identity","whatsapp","gmail","read","message","ask",{"skillNo":92,"name":93,"nameZh":6,"description":94,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":95,"categories":96,"keywords":102,"publishedAt":48},"01M716Y4HF7FVVMJSA2CCVN5VF","hr","Compatibility alias for history-resume; produce the same fresh-state, provenance-aware resume briefing.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fhr\u002FSKILL.md",[97],{"categoryNo":98,"categoryName":99,"taxonomyType":100,"taxonomyTypeDesc":101},"01JBSEEDCAT000000000000038","人力资源","PROFESSION","适用职业",[93,103,104,105,106,40,107,108,109,110,111],"resume","compatibility","provenance","for","produce","alias","briefing","state","fresh",{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":119,"publishedAt":48},"01M716Y1FFNFKFJYYKF8HQA6T8","history-resume","Resume briefing for the current repository branch or worktree using fresh state and admitted history sources.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fhistory-resume\u002FSKILL.md",[118],{"categoryNo":98,"categoryName":99,"taxonomyType":100,"taxonomyTypeDesc":101},[40,103,120,121,39,122,109,123,106,124,125,110,111],"history resume","branch","analysis","repository","worktree","sources",{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":132,"publishedAt":48},"01M716XWXJ8EPHK4RH9T14TR85","grok","Alias for the canonical grokbot browser conversation workflow; \u002Fgrok talks to the Grok bot the same way \u002Fdot talks to Dot.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fgrok\u002FSKILL.md",[],[128,133,134,135,106,136,108,137,138,139],"dot","bot","workflow","canonical","conversation","browser","grokbot",{"skillNo":141,"name":142,"nameZh":6,"description":143,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":144,"categories":145,"keywords":146,"publishedAt":48},"01M70N9Q07DTE0EW5KFXMMMSJ0","codexa-bench","Compare `codexa` and normal `codex` on a fair split of comparable work by default; use `--strict` for duplicate runs of the exact same task.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fcodexa-bench\u002FSKILL.md",[],[147,148,149,37,150,106,151,152,153,154,155],"codexa bench","work","split","default","duplicate","task","compare","normal","codex",{"skillNo":157,"name":158,"nameZh":6,"description":159,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":160,"licenseFileUrl":161,"sourceHtmlUrl":162,"categories":163,"keywords":173,"publishedAt":48},"01M6C24S2X60SP8BBFGAEKPEGW","frontend-design","Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML\u002FCSS layouts, or when styling\u002Fbeautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ffrontend-design\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ffrontend-design\u002FSKILL.md",[164,167,170],{"categoryNo":165,"categoryName":166,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":168,"categoryName":169,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000015","视觉设计",{"categoryNo":171,"categoryName":172,"taxonomyType":100,"taxonomyTypeDesc":101},"01JBSEEDCAT000000000000032","设计师",[174,175,176,37,177,178,179,39,180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197,27,198],"frontend","design","frontend design","generic","ui design","react","html","components","code","artifacts","ui","aesthetics","Production","web components","landing","applications","quality","web","css","create","examples","with","react components","web ui","pages"]