[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79D2CTEEFKBAY3RB5DRHHRK":3,"skill-same:01M79D2CTEEFKBAY3RB5DRHHRK":47},{"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":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":29,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M79D2CTEEFKBAY3RB5DRHHRK","delegation",null,"Decide when to spawn a subagent vs. do work inline (cost-tiered) — don't spawn for trivial work; do spawn for parallel independent investigation or isolated long tasks; use wait for external callbacks and hook_register for external-system handoffs.","PersonalClaw\u002FPersonalClaw","PersonalClaw","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F306824595?v=4","01M79D0KNQCPQXVH30C5PREVCW","01M79D0KPT5XDNEAW87TEYQW3V",9,"MIT","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Fsrc\u002Fpersonalclaw\u002Fskills\u002Fbundled\u002Fdelegation\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: delegation\ndescription: Decide when to spawn a subagent vs. do work inline (cost-tiered) — don't spawn for trivial work; do spawn for parallel independent investigation or isolated long tasks; use wait for external callbacks and hook_register for external-system handoffs.\nalways: false\ntriggers: subagent, spawn, delegate, parallel, in parallel, background task, fan out, isolate, offload, wait for, callback, hook, handoff\n---\n\n# Delegation\n\nSpawning a subagent has real cost — a fresh context, its own token budget, setup\nand synthesis overhead. Delegate when it **pays for itself** in parallelism or\nisolation, and do the work **inline** when it doesn't. Match the mechanism to the\nshape of the work.\n\nTools: `subagent_run` (spawn a worker), `subagent_list` \u002F `subagent_status`\n(track them), `wait` (block for an external signal), `hook_register` (hand off to\nan external system).\n\n## Do it inline (don't spawn)\n\nDefault to inline. **Don't spawn a subagent for trivial or sequential work** that\nshares your context:\n\n- A quick lookup, a single file read\u002Fedit, a one-shot command.\n- Anything that needs your current conversation context to make sense.\n- Sequential steps where each depends on the last — there's nothing to\n  parallelize, and a subagent just adds overhead and a context hand-off.\n\nIf spawning would cost more (setup + synthesis) than just doing it, do it.\n\n## Spawn a subagent (`subagent_run`)\n\nDelegate when the work is either **parallel** or **isolatable**:\n\n- **Parallel independent investigation** — several sub-questions or candidate\n  causes with no ordering dependency and no shared mutable state. Fan them out at\n  once (e.g. probe three hypotheses, research several topics) and synthesize the\n  returns. This is the biggest win: wall-clock time drops.\n- **Isolated long task** — a self-contained chunk of work that would otherwise\n  flood your context with intermediate detail (a deep search, a long grind, a\n  bounded build). Hand it off with a clear deliverable so only the result comes\n  back, keeping your own context lean.\n\nGive each subagent a **self-contained task and an explicit deliverable** — it\ndoesn't share your conversation, so spell out what \"done\" returns. Track them\nwith `subagent_list` \u002F `subagent_status`. **Don't** fan out units that touch the\nsame mutable state — parallel workers racing on the same files corrupt each\nother; parallelize only genuinely independent work.\n\n(For an open-ended grind toward a goal — \"keep working until done\" — a goal loop\nfits better than a raw subagent; see `loop-worker` and `task-and-project`.)\n\n## Waiting on the outside world (`wait`)\n\nWhen the next step depends on something **external** that will arrive later — an\nasync job finishing, an approval, an out-of-band reply — use `wait` to block for\nthat signal instead of busy-polling or spawning a worker whose only job is to\nsleep. `wait` is for \"pause until the callback comes\", not for parallel compute.\n\n## External-system handoffs (`hook_register`)\n\nWhen an **external system** should drive future work — an inbound webhook, an\nevent from another service, a deferred trigger that fires outside this session —\nregister a hook with `hook_register` so PersonalClaw responds when the event\narrives. This is the handoff for \"when X happens out there, do Y here,\" distinct\nfrom spawning a worker now (`subagent_run`) or blocking on one specific signal\n(`wait`).\n\n## Decision summary\n\n- Trivial \u002F sequential \u002F needs my context → **inline**.\n- Independent units, want them concurrent → **`subagent_run`** (parallel).\n- Self-contained long task polluting my context → **`subagent_run`** (isolate).\n- Open-ended grind toward a definition of done → **goal loop**.\n- Pause until an external signal\u002Fcallback → **`wait`**.\n- Let an external system trigger future work → **`hook_register`**.\n","# Delegation\n\nSpawning a subagent has real cost — a fresh context, its own token budget, setup\nand synthesis overhead. Delegate when it **pays for itself** in parallelism or\nisolation, and do the work **inline** when it doesn't. Match the mechanism to the\nshape of the work.\n\nTools: `subagent_run` (spawn a worker), `subagent_list` \u002F `subagent_status`\n(track them), `wait` (block for an external signal), `hook_register` (hand off to\nan external system).\n\n## Do it inline (don't spawn)\n\nDefault to inline. **Don't spawn a subagent for trivial or sequential work** that\nshares your context:\n\n- A quick lookup, a single file read\u002Fedit, a one-shot command.\n- Anything that needs your current conversation context to make sense.\n- Sequential steps where each depends on the last — there's nothing to\n  parallelize, and a subagent just adds overhead and a context hand-off.\n\nIf spawning would cost more (setup + synthesis) than just doing it, do it.\n\n## Spawn a subagent (`subagent_run`)\n\nDelegate when the work is either **parallel** or **isolatable**:\n\n- **Parallel independent investigation** — several sub-questions or candidate\n  causes with no ordering dependency and no shared mutable state. Fan them out at\n  once (e.g. probe three hypotheses, research several topics) and synthesize the\n  returns. This is the biggest win: wall-clock time drops.\n- **Isolated long task** — a self-contained chunk of work that would otherwise\n  flood your context with intermediate detail (a deep search, a long grind, a\n  bounded build). Hand it off with a clear deliverable so only the result comes\n  back, keeping your own context lean.\n\nGive each subagent a **self-contained task and an explicit deliverable** — it\ndoesn't share your conversation, so spell out what \"done\" returns. Track them\nwith `subagent_list` \u002F `subagent_status`. **Don't** fan out units that touch the\nsame mutable state — parallel workers racing on the same files corrupt each\nother; parallelize only genuinely independent work.\n\n(For an open-ended grind toward a goal — \"keep working until done\" — a goal loop\nfits better than a raw subagent; see `loop-worker` and `task-and-project`.)\n\n## Waiting on the outside world (`wait`)\n\nWhen the next step depends on something **external** that will arrive later — an\nasync job finishing, an approval, an out-of-band reply — use `wait` to block for\nthat signal instead of busy-polling or spawning a worker whose only job is to\nsleep. `wait` is for \"pause until the callback comes\", not for parallel compute.\n\n## External-system handoffs (`hook_register`)\n\nWhen an **external system** should drive future work — an inbound webhook, an\nevent from another service, a deferred trigger that fires outside this session —\nregister a hook with `hook_register` so PersonalClaw responds when the event\narrives. This is the handoff for \"when X happens out there, do Y here,\" distinct\nfrom spawning a worker now (`subagent_run`) or blocking on one specific signal\n(`wait`).\n\n## Decision summary\n\n- Trivial \u002F sequential \u002F needs my context → **inline**.\n- Independent units, want them concurrent → **`subagent_run`** (parallel).\n- Self-contained long task polluting my context → **`subagent_run`** (isolate).\n- Open-ended grind toward a definition of done → **goal loop**.\n- Pause until an external signal\u002Fcallback → **`wait`**.\n- Let an external system trigger future work → **`hook_register`**.\n",{"name":5,"always":19,"triggers":23,"description":7},"subagent, spawn, delegate, parallel, in parallel, background task, fan out, isolate, offload, wait for, callback, hook, handoff","main","cfba933d376f99978247a7d18b592232c649e87a","src\u002Fpersonalclaw\u002Fskills\u002Fbundled\u002Fdelegation",6,"https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw","https:\u002F\u002Fpersonalclaw.dev","https:\u002F\u002Fgithub.com\u002FPersonalClaw","Self-hosted personal AI agent — an agentic OS for one person: chat, goal loops, memory, knowledge, automation, apps. Local-first, zero telemetry, MIT.","2026-11-15 04:04:52",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add PersonalClaw\u002FPersonalClaw --skill delegation",{"tool":39,"command":40},"bunx","bunx skills add PersonalClaw\u002FPersonalClaw --skill delegation",{"tool":42,"command":43},"pnpm","pnpm dlx skills add PersonalClaw\u002FPersonalClaw --skill delegation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Fsrc\u002Fpersonalclaw\u002Fskills\u002Fbundled\u002Fdelegation\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":18},"01M79D5S4EBPTFGMRDY7AM7ERD","ledger-read","Read a ledger CSV","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fexternal_formats_golden\u002Fskill-md\u002Fskills\u002Fledger-read\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":18},"01M79D5QFEBGM2PTM21WVTA90K","budget-coach","Weekly spend review","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fexternal_formats_golden\u002Fskill-md\u002Fskills\u002Fbudget-coach\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":18},"01M79D5MTEV62YBKTNY874Y9C4","yt-transcript","Fetch the transcript of a YouTube video and summarise it with timestamps. Use when the user shares a youtube.com or youtu.be link and asks what it says.","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fagent_tool_skills\u002Fyt-transcript\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":18},"01M79D5JG2M47DS0X3GRZ94PXY","incident-writeup","Turn my incident notes, Slack excerpts and graphs into a blameless postmortem in Cartwheel's format. Use after an on-call incident once the mitigation is in.","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fagent_tool_skills\u002Fincident-writeup\u002FSKILL.md",[],[],{"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":18},"01M79D5DX9SWDFDAFW3AD5FATT","feedsmith-release","Cut a feedsmith release - bump the version, move the Unreleased changelog section, tag, and draft the GitHub release notes. Use when I say \"release feedsmith\" or \"cut 0.x\".","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fagent_tool_skills\u002Ffeedsmith-release\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":18},"01M79D59Z7A0FEZSB5AEQYHNSR","feedsmith-bench","Benchmark feedsmith fetch against a real OPML, cold and warm, and compare with main. Use before merging anything that touches fetch.py or store.py.","https:\u002F\u002Fgithub.com\u002FPersonalClaw\u002FPersonalClaw\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fagent_tool_homes\u002Fnoor\u002F.codex\u002Fskills\u002Ffeedsmith-bench\u002FSKILL.md",[],[]]