{"id":"ed183c51-03ad-44d3-b156-b94c036aceef","shortId":"tFqjww","kind":"skill","title":"gsd-coordinator","tagline":"End-to-end task orchestration across AI coding engines. Coordinates Claude, Codex, Codex Spark (the faster Codex variant), and OpenCode workers using dispatch-verify-synthesize workflows. Requires the upstream agent-mux CLI for worker dispatch. Use when work spans multiple depend","description":"**What this skill does:** The GSD Coordinator manages complex, multi-step tasks by dispatching work to different AI engines (Claude, Codex, OpenCode) and combining their results. Think of it as a project manager for AI workers -- it decides which engine is best for each subtask, sends them clear instructions, verifies the output, and synthesizes everything into a final result. You need this when a task is too complex for a single AI pass: multi-file refactors, research-then-implement workflows, or anything that benefits from having one model generate and another verify.\n\n# GSD Coordinator\n\nYou are a GSD (Get Shit Done) coordinator. Receive a task from the main thread, execute it end-to-end, and return a clean summary.\n\nYou have all standard tools (Read, Write, Edit, Bash, Grep, Glob, WebFetch, WebSearch). Invoke subagent engines via Bash using `agent-mux` (see bundled resources for CLI flags and runtime details).\n\nKey constraint: do not spawn Claude subagents via `Task`; use `agent-mux --engine claude` instead. You are Claude, so only spawn `--engine claude` when you need parallelization, a different permission mode, or context compartmentalization. Use Codex for model diversity.\n\n## Prerequisites\n\n- Install `agent-mux` CLI: https://github.com/buildoak/agent-mux\n- Installation docs: https://github.com/buildoak/agent-mux#readme\n- Works with Claude Code and Codex CLI\n\n## Dependency Boundaries\n\n- Required external dependency: upstream `agent-mux` CLI (`https://github.com/buildoak/agent-mux`).\n- Optional local dependency: `skills/agent-mux` in this repo is an integration pointer only, not required if `agent-mux` is already installed.\n- Optional skill dependencies: per-worker domain skills are injected only when useful (`--skill foo`); they are not required for coordinator baseline operation.\n\n## Setup\n\n```bash\n# Install agent-mux CLI (required upstream dependency)\ngit clone https://github.com/buildoak/agent-mux.git /path/to/agent-mux\ncd /path/to/agent-mux && ./setup.sh && bun link\n```\n\n- **Claude Code:** copy this skill folder into `.claude/skills/gsd-coordinator/`\n- **Codex CLI:** append this SKILL.md content to your project's root `AGENTS.md`\n\nFor the full installation walkthrough (prerequisites, verification, troubleshooting), see [references/installation-guide.md](references/installation-guide.md).\n\n### Artifact directory setup\n\nWorkers occasionally write file artifacts (outputs exceeding 200 lines, deliverables). By default, these go to `_workbench/` inside this skill directory. A `.gitkeep` ships with the skill to create it.\n\nIf you prefer a different location -- a project scratch folder, a `tmp/` directory, wherever -- set the path when you first configure the skill. The SKILL.md uses `<artifacts-dir>` as a placeholder throughout. Replace it with your chosen path, or leave the default:\n\n```text\nDefault:  .claude/skills/gsd-coordinator/_workbench/\nOverride: any directory you prefer for working artifacts\n```\n\nThe coordinator will write artifacts to `<artifacts-dir>/YYYY-MM-DD-{engine}-{description}.md`. Deliverables (final outputs) always go to their canonical destination, not the artifacts directory.\n\n## Staying Updated\n\nThis skill ships with an `UPDATES.md` changelog and `UPDATE-GUIDE.md` for your AI agent.\n\nAfter installing, tell your agent: \"Check `UPDATES.md` in the gsd-coordinator skill for any new features or changes.\"\n\nWhen updating, tell your agent: \"Read `UPDATE-GUIDE.md` and apply the latest changes from `UPDATES.md`.\"\n\nFollow `UPDATE-GUIDE.md` so customized local files are diffed before any overwrite.\n\n## Quick Start\n\nRun this checklist before first orchestration:\n\n1. Verify CLI is installed: `agent-mux --help`\n2. Run one bounded worker task:\n   ```bash\n   agent-mux --engine codex --reasoning high \"List changed files and summarize intent in 5 bullets\"\n   ```\n3. Parse the JSON output and keep only the worker's `response` field.\n4. Escalate to multi-step orchestration only after the single worker pass succeeds.\n\n## Two Operating Modes\n\n### 1. Claude Code mode (full orchestration)\n\nUse when the coordinator can dispatch workers and close the loop directly.\n\n- Can orchestrate the full loop: dispatch -> verify -> synthesize -> fix -> re-dispatch\n- Can use `agent-mux` for Codex, Codex Spark, Claude, and OpenCode workers\n- Can read/write files and run verification commands\n\n### 2. Codex mode (planner only)\n\nUse when `gsd-coordinator` is loaded by a Codex worker that cannot orchestrate nested workers directly.\n\n- Reads this runbook as a decision framework\n- Selects pattern, workers, prompts, and artifacts\n- Returns an execution plan instead of dispatching nested workers\n- Must return `status: needs-orchestrator` so a parent orchestrator executes the plan\n\n## Know Your Workers\n\nMatch the right worker to each step:\n\n- Claude Opus 4.6 (`--engine claude`): Natural orchestrator. Thrives on ambiguity, decides from available info, and writes strong prompts. Use for architecture, synthesis, open-ended exploration, and prompt crafting.\n- Codex 5.3 (`--engine codex`): Precise executor. Pedantic, thorough, and detail-attentive. Needs explicit scope: one goal, specific files, explicit output path. Use `--reasoning high` (standard deep-reasoning mode) for implementation; reserve `xhigh` (maximum reasoning mode) for deep audits.\n- Codex Spark (`--engine codex --model gpt-5.3-codex-spark`): Same precision style, much faster, smaller context window. Use for parallel workers, filesystem scanning, and focused medium-difficulty tasks.\n- OpenCode (`--engine opencode`): Supplemental third-opinion engine for model-lineage diversity.\n  - `kimi`\n  - `glm-5`\n  - `opencode-minimax`\n  - `free`\n\n## Default Playbook\n\nFor any task, follow this sequence. Deviate when you have a reason.\n\n1. Triage: read the task and identify inputs, outputs, and constraints.\n2. Pick a pattern: Implementation, Audit, Research, or Fan-Out (parallel independent workers).\n3. Select skills: for each step, identify which skills (if any) the worker needs.\n4. Choose workers: match each step to the right engine.\n5. Write prompt specs: one goal, specific files, explicit output path.\n6. Run: execute workers with skills injected and parse output to extract the response field.\n7. Verify: read artifacts, check quality, fix or re-run if needed.\n8. Return: write the primary artifact, compose summary, report status.\n\n## Model Selection Heuristics\n\n### The Core Question: What Does This Step Need?\n\n| Step needs... | Use | Why |\n|---|---|---|\n| Exploration, ambiguity resolution | Claude | Codex flounders without scope |\n| Precise implementation | Codex (`high`) | Pedantic, detail-oriented |\n| Deep architecture audit | Codex (`xhigh`) | Catches edge cases `high` misses |\n| Fast parallel grunt work | Codex Spark | Speed and focus on bounded tasks |\n| Synthesis or documentation | Claude | Strong structured output |\n| Third-opinion verification | OpenCode | Independent lineage cross-check |\n\n### Fan-Out vs Go Deep\n\nFan out (parallel workers) when subtasks are independent, speed matters, tasks are medium difficulty, or you need broad coverage.\n\nGo deep (single worker, `high`/`xhigh`) when multi-file reasoning is required, context pressure is high, the task already failed once, or getting it wrong has high downstream cost.\n\n### Escalation Heuristic\n\nStart at `high`. If wrong or incomplete, escalate to `xhigh`. If `xhigh` also fails, fix the prompt and decomposition; do not retry blindly.\n\n## Orchestration Patterns\n\n### 10x Pattern (Codex Generate + Opus Audit)\n\nMost validated pipeline. Different blind spots produce higher confidence.\n\n1. Spawn Codex at `high` to generate or refactor.\n2. Read its output yourself (you are Opus).\n3. Fix issues or spawn another Codex pass.\n4. Write the final artifact and return summary.\n\nUse when implementation, code review, or mechanical refactoring is needed. Skip when inline work is faster or the task is purely writing/synthesis.\n\n### Fan-Out\n\nSpawn N parallel workers on independent subtasks.\n\nWorkers return inline by default. If output exceeds 200 lines, write to `<artifacts-dir>/YYYY-MM-DD-{engine}-{topic}.md`.\n\nRead all worker outputs and synthesize into one final output.\n\n### Research + Synthesize\n\nRead relevant files and references. Use web search if needed. Synthesize into one artifact and return summary.\n\n### Triple-Check\n\nUse three model lineages for high-stakes work.\n\n1. Claude frames/decomposes approach.\n2. Codex implements or performs core verification.\n3. OpenCode provides independent verification.\n\n## Bring Your Own Skills\n\nSkills are operational blueprints. Each skill `SKILL.md` bundles domain knowledge, conventions, and CLI workflows into a reusable playbook.\n\nWhen dispatching a worker with `agent-mux`, inject only the relevant skill(s):\n\n```bash\nagent-mux --engine codex --skill your-skill --reasoning high \"Search for auth architecture docs\"\n```\n\n```bash\nagent-mux --engine codex --skill your-read-skill --skill your-write-skill --reasoning high \"Read the existing spec, then write the updated version\"\n```\n\nRules:\n\n- A skill-equipped worker should follow the skill playbook, not ad-hoc reasoning.\n- Do not over-inject. Pick only skills needed for that subtask.\n- If no skill fits, prompt the worker directly. Skills are an accelerator, not a requirement.\n\n## Output Contract\n\nDefault: return inline.\n\nWorkers return focused summaries inline by default. Do not write file artifacts unless output exceeds 200 lines or a deliverable is explicitly required.\n\nWhen files are needed:\n\n- Over 200 lines: write to `<artifacts-dir>/YYYY-MM-DD-{engine}-{description}.md` (defaults to `_workbench/` inside this skill directory; see [Artifact directory setup](#artifact-directory-setup))\n- Deliverables: write directly to final destination (for example `<your-project>/research/` or other project folders)\n\nNaming when files are written: `YYYY-MM-DD-{engine}-{description}.md`\n\n- `engine` = `codex` | `claude` | `spark` | `opencode` | `coordinator`\n- `description` = descriptive kebab-case\n- Parallel workers: add suffixes like `YYYY-MM-DD-spark-topic-a.md` and `YYYY-MM-DD-spark-topic-b.md`\n\nMinimal frontmatter when files are written:\n\n- `date: YYYY-MM-DD`\n- `engine: codex | claude | spark | opencode | coordinator`\n- `status: complete | partial | error`\n\nSandbox rule: Codex workers that write files must use `--sandbox workspace-write --cwd <repo-root>`.\n\nFormat Cards: when writing to canonical locations (not `<artifacts-dir>/`), use the correct frontmatter required by your project's conventions. Do not invent domains.\n\n## Context Discipline\n\n- Workers return briefs, not dumps. Ask for focused summaries like \"Return a 3-5 sentence summary\" or \"Return the file path and a one-paragraph verdict.\"\n- Pass paths, not content. Hand off file paths between steps; the next worker reads what it needs.\n- Check background workers with `tail -n 20` via Bash. Do not load full long output files into context.\n- Scope your reads. Verify specific sections, not whole files. Use offset/limit or Grep when possible.\n\n## Return Contract\n\nWhen finished, return to the main thread:\n\n1. File path to the primary artifact (if any)\n2. 3-5 sentence summary of work, findings, and decisions\n3. Status: `done` | `blocked` | `needs-decision` | `needs-orchestrator` (planner mode only)\n\nNever dump raw content back. Always return path + summary + status.\n\n## Error Handling\n\n| Symptom | Cause | Recovery |\n|---------|-------|----------|\n| `agent-mux: command not found` | Upstream CLI not installed or missing from `$PATH` | Install from `https://github.com/buildoak/agent-mux` and re-run with absolute path if needed |\n| Worker returns malformed/non-JSON output | Wrong runtime flags or worker crash | Re-run with strict JSON mode in `agent-mux`, capture stderr, and reduce prompt scope |\n| Worker output is empty or generic | Prompt is underspecified | Rewrite prompt with one goal, explicit files, and expected output contract |\n| Repeated failures on same step | Wrong engine or decomposition | Switch engine (Codex vs Claude), split the step, and re-dispatch |\n| Coordinator cannot orchestrate nested workers | Running in planner-only environment | Return `status: needs-orchestrator` with an executable plan |\n\n## Anti-Patterns\n\n| Do NOT | Do instead |\n|--------|------------|\n| Blind retry worker failures | Diagnose root cause (engine choice, scope, skill, prompt quality) before retrying |\n| Paste full artifacts into prompts | Write artifacts to files and pass only file paths |\n| Send exploration to Codex or focused implementation to Claude by habit | Match worker to step needs with the heuristics table |\n| Spawn Claude for more Claude by default | Use Codex/OpenCode for model diversity unless parallelization/context isolation is needed |\n| Use `xhigh` for routine work | Default to `high`; escalate only after failure or high-stakes risk |\n| Assume main-thread context is present | Re-read required files in each worker prompt |\n| Skip skill injection when a relevant skill exists | Inject only the minimal relevant skill set |\n| Over-prompt with long novels | Give one goal, relevant files, and expected output format |\n| Run Triple-Check for low-risk tasks | Use proportional verification based on risk |\n\n## Bundled Resources Index\n\n| Path | What | When to load |\n|------|------|--------------|\n| `./UPDATES.md` | Structured changelog for AI agents | When checking for new features or updates |\n| `./UPDATE-GUIDE.md` | Instructions for AI agents performing updates | When updating this skill |\n| `./references/installation-guide.md` | Detailed install walkthrough for Claude Code and Codex CLI | First-time setup or environment repair |\n| `./references/orchestration-examples.md` | Prompting and orchestration examples | When building or debugging dispatch plans |\n| `https://github.com/buildoak/agent-mux` | Required upstream CLI docs and runtime contract | Always for installation/flags/runtime behavior |","tags":["gsd","coordinator","fieldwork","skills","buildoak","agent-skills","ai-agents","ai-tools","automation","browser-automation","claude-code","claude-skills"],"capabilities":["skill","source-buildoak","skill-gsd-coordinator","topic-agent-skills","topic-ai-agents","topic-ai-tools","topic-automation","topic-browser-automation","topic-claude-code","topic-claude-skills","topic-codex"],"categories":["fieldwork-skills"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/buildoak/fieldwork-skills/gsd-coordinator","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add buildoak/fieldwork-skills","source_repo":"https://github.com/buildoak/fieldwork-skills","install_from":"skills.sh"}},"qualityScore":"0.457","qualityRationale":"deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 15 github stars · SKILL.md body (13,948 chars)","verified":false,"liveness":"unknown","lastLivenessCheck":null,"agentReviews":{"count":0,"score_avg":null,"cost_usd_avg":null,"success_rate":null,"latency_p50_ms":null,"narrative_summary":null,"summary_updated_at":null},"enrichmentModel":"deterministic:skill-github:v1","enrichmentVersion":1,"enrichedAt":"2026-04-22T19:06:32.949Z","embedding":null,"createdAt":"2026-04-18T23:07:13.978Z","updatedAt":"2026-04-22T19:06:32.949Z","lastSeenAt":"2026-04-22T19:06:32.949Z","tsv":"'-5':838,1555,1639 '-5.3':799 '/buildoak/agent-mux':250,275,1693,2004 '/buildoak/agent-mux#readme':255 '/buildoak/agent-mux.git':334 '/path/to/agent-mux':335,337 '/references/installation-guide.md':1974 '/references/orchestration-examples.md':1991 '/research':1447 '/setup.sh':338 '/update-guide.md':1963 '/updates.md':1950 '/yyyy-mm-dd-':461,1199,1420 '1':545,607,857,1122,1244,1628 '10x':1107 '2':554,657,868,1131,1248,1637 '20':1592 '200':382,1195,1403,1416 '3':577,882,1139,1255,1554,1638,1647 '4':590,896,1147 '4.6':726 '5':575,906 '5.3':754 '6':917 '7':932 '8':945 'absolut':1699 'acceler':1379 'across':10 'ad':1353 'ad-hoc':1352 'add':1477 'agent':36,191,213,245,270,292,324,492,497,516,551,562,640,1288,1298,1315,1676,1722,1955,1967 'agent-mux':35,190,212,244,269,291,323,550,561,639,1287,1297,1314,1675,1721 'agents.md':360 'ai':11,66,83,120,491,1954,1966 'alreadi':295,1069 'also':1094 'alway':468,1665,2012 'ambigu':733,971 'anoth':141,1144 'anti':1792 'anti-pattern':1791 'anyth':132 'append':351 'appli':520 'approach':1247 'architectur':744,987,1311 'artifact':372,379,454,459,476,691,935,950,1151,1228,1399,1432,1436,1634,1815,1819 'artifact-directory-setup':1435 'ask':1547 'assum':1881 'attent':764 'audit':792,873,988,1112 'auth':1310 'avail':736 'back':1664 'background':1587 'base':1939 'baselin':318 'bash':179,188,321,560,1296,1313,1594 'behavior':2015 'benefit':134 'best':90 'blind':1104,1117,1798 'block':1650 'blueprint':1267 'bound':557,1006 'boundari':264 'brief':1544 'bring':1260 'broad':1048 'build':1997 'bullet':576 'bun':339 'bundl':194,1271,1942 'cannot':674,1772 'canon':472,1523 'captur':1724 'card':1519 'case':993,1474 'catch':991 'caus':1673,1804 'cd':336 'chang':511,523,569 'changelog':486,1952 'check':498,936,1024,1234,1586,1930,1957 'checklist':541 'choic':1806 'choos':897 'chosen':438 'claud':15,68,207,216,220,225,258,341,608,646,724,728,973,1011,1245,1466,1496,1763,1835,1848,1851,1979 'claude/skills/gsd-coordinator':348 'claude/skills/gsd-coordinator/_workbench':446 'clean':169 'clear':96 'cli':38,197,247,262,272,326,350,547,1276,1682,1983,2007 'clone':331 'close':621 'code':12,259,342,609,1158,1980 'codex':16,17,21,69,238,261,349,565,643,644,658,671,753,756,793,796,801,974,980,989,1000,1109,1124,1145,1249,1301,1318,1465,1495,1506,1761,1830,1982 'codex-spark':800 'codex/opencode':1855 'combin':72 'command':656,1678 'compartment':236 'complet':1501 'complex':56,116 'compos':951 'confid':1121 'configur':424 'constraint':203,867 'content':354,1572,1663 'context':235,809,1063,1540,1603,1885 'contract':1384,1620,1749,2011 'convent':1274,1535 'coordin':3,14,54,144,152,317,456,504,616,666,1469,1499,1771 'copi':343 'core':959,1253 'correct':1528 'cost':1079 'coverag':1049 'craft':752 'crash':1712 'creat':402 'cross':1023 'cross-check':1022 'custom':529 'cwd':1517 'date':1489 'dd':1460,1493 'debug':1999 'decid':86,734 'decis':684,1646,1653 'decomposit':1100,1758 'deep':780,791,986,1030,1051 'deep-reason':779 'default':386,443,445,843,1191,1385,1394,1424,1853,1869 'deliver':384,465,1407,1439 'depend':47,263,267,278,299,329 'descript':463,1422,1462,1470,1471 'destin':473,1444 'detail':201,763,984,1975 'detail-attent':762 'detail-ori':983 'deviat':851 'diagnos':1802 'dif':533 'differ':65,231,408,1116 'difficulti':821,1044 'direct':624,678,1375,1441 'directori':373,394,416,449,477,1430,1433,1437 'disciplin':1541 'dispatch':28,41,62,618,630,636,698,1283,1770,2000 'dispatch-verify-synthes':27 'divers':241,835,1858 'doc':252,1312,2008 'document':1010 'domain':303,1272,1539 'done':151,1649 'downstream':1078 'dump':1546,1661 'edg':992 'edit':178 'empti':1733 'end':5,7,163,165,748 'end-to-end':4,162 'engin':13,67,88,186,215,224,462,564,727,755,795,824,830,905,1200,1300,1317,1421,1461,1464,1494,1756,1760,1805 'environ':1781,1989 'equip':1344 'error':1503,1670 'escal':591,1080,1089,1872 'everyth':103 'exampl':1446,1995 'exceed':381,1194,1402 'execut':160,694,711,919,1789 'executor':758 'exist':1333,1904 'expect':1747,1924 'explicit':766,772,914,1409,1744 'explor':749,970,1828 'extern':266 'extract':928 'fail':1070,1095 'failur':1751,1801,1875 'fan':877,1026,1031,1178 'fan-out':876,1025,1177 'fast':996 'faster':20,807,1170 'featur':509,1960 'field':589,931 'file':124,378,531,570,652,771,913,1059,1217,1398,1412,1454,1486,1510,1561,1575,1601,1612,1629,1745,1821,1825,1892,1922 'filesystem':815 'final':106,466,1150,1211,1443 'find':1644 'finish':1622 'first':423,543,1985 'first-tim':1984 'fit':1371 'fix':633,938,1096,1140 'flag':198,1709 'flounder':975 'focus':818,1004,1390,1549,1832 'folder':346,413,1451 'follow':526,848,1347 'foo':311 'format':1518,1926 'found':1680 'frames/decomposes':1246 'framework':685 'free':842 'frontmatt':1484,1529 'full':363,611,628,1598,1814 'generat':139,1110,1128 'generic':1735 'get':149,1073 'git':330 'github.com':249,254,274,333,1692,2003 'github.com/buildoak/agent-mux':248,273,1691,2002 'github.com/buildoak/agent-mux#readme':253 'github.com/buildoak/agent-mux.git':332 'gitkeep':396 'give':1918 'glm':837 'glob':181 'go':388,469,1029,1050 'goal':769,911,1743,1920 'gpt':798 'grep':180,1616 'grunt':998 'gsd':2,53,143,148,503,665 'gsd-coordin':1,502,664 'habit':1837 'hand':1573 'handl':1671 'help':553 'heurist':957,1081,1845 'high':567,777,981,994,1054,1066,1077,1084,1126,1241,1307,1330,1871,1878 'high-stak':1240,1877 'higher':1120 'hoc':1354 'identifi':863,888 'implement':129,784,872,979,1157,1250,1833 'incomplet':1088 'independ':880,1020,1038,1185,1258 'index':1944 'info':737 'inject':306,923,1290,1360,1899,1905 'inlin':1167,1189,1387,1392 'input':864 'insid':391,1427 'instal':243,251,296,322,364,494,549,1684,1689,1976 'installation/flags/runtime':2014 'instead':217,696,1797 'instruct':97,1964 'integr':285 'intent':573 'invent':1538 'invok':184 'isol':1861 'issu':1141 'json':580,1718 'kebab':1473 'kebab-cas':1472 'keep':583 'key':202 'kimi':836 'know':714 'knowledg':1273 'latest':522 'leav':441 'like':1479,1551 'line':383,1196,1404,1417 'lineag':834,1021,1238 'link':340 'list':568 'load':668,1597,1949 'local':277,530 'locat':409,1524 'long':1599,1916 'loop':623,629 'low':1933 'low-risk':1932 'main':158,1626,1883 'main-thread':1882 'malformed/non-json':1705 'manag':55,81 'match':717,899,1838 'matter':1040 'maximum':787 'md':464,1202,1423,1463 'mechan':1161 'medium':820,1043 'medium-difficulti':819 'minim':1483,1908 'minimax':841 'miss':995,1686 'mm':1459,1492 'mode':233,606,610,659,782,789,1658,1719 'model':138,240,797,833,955,1237,1857 'model-lineag':832 'much':806 'multi':58,123,594,1058 'multi-fil':122,1057 'multi-step':57,593 'multipl':46 'must':701,1511 'mux':37,192,214,246,271,293,325,552,563,641,1289,1299,1316,1677,1723 'n':1181,1591 'name':1452 'natur':729 'need':109,228,705,765,895,944,965,967,1047,1164,1224,1364,1414,1585,1652,1655,1702,1785,1842,1863 'needs-decis':1651 'needs-orchestr':704,1654,1784 'nest':676,699,1774 'never':1660 'new':508,1959 'next':1580 'novel':1917 'occasion':376 'offset/limit':1614 'one':137,556,768,910,1210,1227,1566,1742,1919 'one-paragraph':1565 'open':747 'open-end':746 'opencod':24,70,648,823,825,840,1019,1256,1468,1498 'opencode-minimax':839 'oper':319,605,1266 'opinion':829,1017 'option':276,297 'opus':725,1111,1138 'orchestr':9,544,596,612,626,675,706,710,730,1105,1656,1773,1786,1994 'orient':985 'output':100,380,467,581,773,865,915,926,1014,1134,1193,1206,1212,1383,1401,1600,1706,1731,1748,1925 'over-inject':1358 'over-prompt':1912 'overrid':447 'overwrit':536 'paragraph':1567 'parallel':229,813,879,997,1033,1182,1475 'parallelization/context':1860 'parent':709 'pars':578,925 'partial':1502 'pass':121,602,1146,1569,1823 'past':1813 'path':420,439,774,916,1562,1570,1576,1630,1667,1688,1700,1826,1945 'pattern':687,871,1106,1108,1793 'pedant':759,982 'per':301 'per-work':300 'perform':1252,1968 'permiss':232 'pick':869,1361 'pipelin':1115 'placehold':432 'plan':695,713,1790,2001 'planner':660,1657,1779 'planner-on':1778 'playbook':844,1281,1350 'pointer':286 'possibl':1618 'precis':757,804,978 'prefer':406,451 'prerequisit':242,366 'present':1887 'pressur':1064 'primari':949,1633 'produc':1119 'project':80,357,411,1450,1533 'prompt':689,741,751,908,1098,1372,1728,1736,1740,1809,1817,1896,1914,1992 'proport':1937 'provid':1257 'pure':1175 'qualiti':937,1810 'question':960 'quick':537 'raw':1662 're':635,941,1696,1714,1769,1889 're-dispatch':634,1768 're-read':1888 're-run':940,1695,1713 'read':176,517,679,859,934,1132,1203,1215,1322,1331,1582,1606,1890 'read/write':651 'reason':566,776,781,788,856,1060,1306,1329,1355 'receiv':153 'recoveri':1674 'reduc':1727 'refactor':125,1130,1162 'refer':1219 'references/installation-guide.md':370,371 'relev':1216,1293,1902,1909,1921 'repair':1990 'repeat':1750 'replac':434 'repo':282 'report':953 'requir':32,265,289,315,327,1062,1382,1410,1530,1891,2005 'research':127,874,1213 'research-then-impl':126 'reserv':785 'resolut':972 'resourc':195,1943 'respons':588,930 'result':74,107 'retri':1103,1799,1812 'return':167,692,702,946,1153,1188,1230,1386,1389,1543,1552,1559,1619,1623,1666,1704,1782 'reusabl':1280 'review':1159 'rewrit':1739 'right':719,904 'risk':1880,1934,1941 'root':359,1803 'routin':1867 'rule':1340,1505 'run':539,555,654,918,942,1697,1715,1776,1927 'runbook':681 'runtim':200,1708,2010 'sandbox':1504,1513 'scan':816 'scope':767,977,1604,1729,1807 'scratch':412 'search':1222,1308 'section':1609 'see':193,369,1431 'select':686,883,956 'send':94,1827 'sentenc':1556,1640 'sequenc':850 'set':418,1911 'setup':320,374,1434,1438,1987 'ship':397,482 'shit':150 'singl':119,600,1052 'skill':50,298,304,310,345,393,400,426,481,505,884,890,922,1263,1264,1269,1294,1302,1305,1319,1323,1324,1328,1343,1349,1363,1370,1376,1429,1808,1898,1903,1910,1973 'skill-equip':1342 'skill-gsd-coordinator' 'skill.md':353,428,1270 'skills/agent-mux':279 'skip':1165,1897 'smaller':808 'source-buildoak' 'span':45 'spark':18,645,794,802,1001,1467,1497 'spawn':206,223,1123,1143,1180,1847 'spec':909,1334 'specif':770,912,1608 'speed':1002,1039 'split':1764 'spot':1118 'stake':1242,1879 'standard':174,778 'start':538,1082 'status':703,954,1500,1648,1669,1783 'stay':478 'stderr':1725 'step':59,595,723,887,901,964,966,1578,1754,1766,1841 'strict':1717 'strong':740,1012 'structur':1013,1951 'style':805 'subag':185,208 'subtask':93,1036,1186,1367 'succeed':603 'suffix':1478 'summar':572 'summari':170,952,1154,1231,1391,1550,1557,1641,1668 'supplement':826 'switch':1759 'symptom':1672 'synthes':30,102,632,1208,1214,1225 'synthesi':745,1008 'tabl':1846 'tail':1590 'task':8,60,113,155,210,559,822,847,861,1007,1041,1068,1173,1935 'tell':495,514 'text':444 'think':75 'third':828,1016 'third-opinion':827,1015 'thorough':760 'thread':159,1627,1884 'three':1236 'thrive':731 'throughout':433 'time':1986 'tmp':415 'tool':175 'topic':1201 'topic-agent-skills' 'topic-ai-agents' 'topic-ai-tools' 'topic-automation' 'topic-browser-automation' 'topic-claude-code' 'topic-claude-skills' 'topic-codex' 'triag':858 'tripl':1233,1929 'triple-check':1232,1928 'troubleshoot':368 'two':604 'underspecifi':1738 'unless':1400,1859 'updat':479,513,1338,1962,1969,1971 'update-guide.md':488,518,527 'updates.md':485,499,525 'upstream':34,268,328,1681,2006 'use':26,42,189,211,237,309,429,613,638,662,742,775,811,968,1155,1220,1235,1512,1526,1613,1854,1864,1936 'valid':1114 'variant':22 'verdict':1568 'verif':367,655,1018,1254,1259,1938 'verifi':29,98,142,546,631,933,1607 'version':1339 'via':187,209,1593 'vs':1028,1762 'walkthrough':365,1977 'web':1221 'webfetch':182 'websearch':183 'wherev':417 'whole':1611 'window':810 'without':976 'work':44,63,256,453,999,1168,1243,1643,1868 'workbench':390,1426 'worker':25,40,84,302,375,558,586,601,619,649,672,677,688,700,716,720,814,881,894,898,920,1034,1053,1183,1187,1205,1285,1345,1374,1388,1476,1507,1542,1581,1588,1703,1711,1730,1775,1800,1839,1895 'workflow':31,130,1277 'workspac':1515 'workspace-writ':1514 'write':177,377,458,739,907,947,1148,1197,1327,1336,1397,1418,1440,1509,1516,1521,1818 'writing/synthesis':1176 'written':1456,1488 'wrong':1075,1086,1707,1755 'xhigh':786,990,1055,1091,1093,1865 'your-read-skil':1320 'your-skil':1303 'your-write-skil':1325 'yyyi':1458,1491 'yyyy-mm-dd':1457,1490 'yyyy-mm-dd-spark-topic-a.md':1480 'yyyy-mm-dd-spark-topic-b.md':1482","prices":[{"id":"961b4e7c-3b96-491c-9b7d-8ab12720e662","listingId":"ed183c51-03ad-44d3-b156-b94c036aceef","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"buildoak","category":"fieldwork-skills","install_from":"skills.sh"},"createdAt":"2026-04-18T23:07:13.978Z"}],"sources":[{"listingId":"ed183c51-03ad-44d3-b156-b94c036aceef","source":"github","sourceId":"buildoak/fieldwork-skills/gsd-coordinator","sourceUrl":"https://github.com/buildoak/fieldwork-skills/tree/main/skills/gsd-coordinator","isPrimary":false,"firstSeenAt":"2026-04-18T23:07:13.978Z","lastSeenAt":"2026-04-22T19:06:32.949Z"}],"details":{"listingId":"ed183c51-03ad-44d3-b156-b94c036aceef","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"buildoak","slug":"gsd-coordinator","github":{"repo":"buildoak/fieldwork-skills","stars":15,"topics":["agent-skills","ai-agents","ai-tools","automation","browser-automation","claude-code","claude-skills","codex"],"license":"apache-2.0","html_url":"https://github.com/buildoak/fieldwork-skills","pushed_at":"2026-03-18T08:36:25Z","description":"Battle-tested skills for AI agents that do real work","skill_md_sha":"f7a89c15bb720dd1d6de79fd114b44a8d1df39fb","skill_md_path":"skills/gsd-coordinator/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/buildoak/fieldwork-skills/tree/main/skills/gsd-coordinator"},"layout":"multi","source":"github","category":"fieldwork-skills","frontmatter":{"name":"gsd-coordinator","description":"End-to-end task orchestration across AI coding engines. Coordinates Claude, Codex, Codex Spark (the faster Codex variant), and OpenCode workers using dispatch-verify-synthesize workflows. Requires the upstream agent-mux CLI for worker dispatch. Use when work spans multiple dependent steps, benefits from model diversity, or needs structured quality verification."},"skills_sh_url":"https://skills.sh/buildoak/fieldwork-skills/gsd-coordinator"},"updatedAt":"2026-04-22T19:06:32.949Z"}}