[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6GGRBHASN4JCSSCB139F3QT":3,"skill-same:01M6GGRBHASN4JCSSCB139F3QT":67},{"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":23,"publishedAt":41,"linkOnly":42,"markdownContent":43,"markdownBody":44,"frontMatter":45,"refName":46,"commitSha":47,"skillDirectory":48,"forksCount":49,"repositoryHtmlUrl":50,"repositoryHomepageUrl":6,"ownerHtmlUrl":51,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":52,"install":53,"resources":66},"01M6GGRBHASN4JCSSCB139F3QT","subagent-dispatch",null,"When delegating work to a subagent — implementation, review, research, or exploration. Use when a task benefits from a fresh context window, parallel execution, or isolated scope.","ahmadulhoq\u002Fagentskel","ahmadulhoq","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F3884252?v=4","01M6GGC32SQTQ52FY4JB14FYSY","Ahmadul Hoq","01M6GGC37CJEYM63V3WEHY4Y8V",13,"MIT","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsubagent-dispatch\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000021","资料检索与研究","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40],"dispatch","subagent dispatch","subagent","execution","parallel","work","context","use","review","parallel execution","implementation","task","exploration","fresh","scope","benefits","research","2026-11-10 10:12:30",false,"---\nname: subagent-dispatch\nlicense: MIT\ndescription: When delegating work to a subagent — implementation, review, research, or exploration. Use when a task benefits from a fresh context window, parallel execution, or isolated scope.\n---\n\n# Subagent Dispatch\n\n**Principle:** The parent agent has full context (rules, memory, conventions). It\ncreates the plan. Subagents execute plan steps — they don't rediscover the project.\n\n---\n\n## When to Dispatch\n\n| Signal | Why |\n|---|---|\n| Plan has independent steps that can run in parallel | Concurrent execution saves time |\n| Context window is getting large | Fresh subagent avoids attention decay |\n| Research would pollute main context with raw file reads | Subagent returns a summary |\n| Code review needs a fresh perspective | Reviewer hasn't seen implementation reasoning |\n\n**Do NOT dispatch when:** steps have sequential dependencies, or the task is simpler\nto do directly than to explain.\n\n---\n\n## How to Dispatch\n\nThe parent creates the plan first (via task-planner or workflow Phase 1), then\nidentifies which steps can run in parallel.\n\n### Implementation dispatch\nPass the specific plan steps. The parent already checked rules, conventions, and\nsacred behaviors when creating the plan — the subagent doesn't need to re-read them.\n\n```\n\"Execute steps 3-5 from the plan:\n - Step 3: [exact instruction]\n - Step 4: [exact instruction]\n - Step 5: [exact instruction]\n Files to modify: [list].\n Do not commit or push.\"\n```\n\n### Review dispatch\nThe reviewer needs to know what to check against. Include references to constraints.\n\n```\n\"Review the changes made to [files].\n Read .memory\u002FSACRED.md — verify no sacred behaviors were modified.\n Read .memory\u002FCONVENTIONS.md — verify code follows project conventions.\n Check against the plan: [paste or summarize the plan].\n Report: issues found, verdict (approve\u002Frequest changes).\"\n```\n\n### Research dispatch\nPass the question. No rules needed.\n\n```\n\"How does [module X] connect to [module Y]? Read MAP.md for orientation.\n Report: answer, evidence (file:line), relevant files.\"\n```\n\n---\n\n## Validate Results\n\n- [ ] Read the subagent's output\n- [ ] Verify it meets the plan's success criteria\n- [ ] Check for side effects (unexpected file changes)\n- [ ] If validation fails, re-dispatch with corrected instructions\n\n---\n\n## Rules\n\n1. **Subagents do NOT run session-start.** They execute plan steps, not session init.\n2. **Subagents do NOT commit or push.** The parent handles git.\n3. **One task per subagent.** Don't batch unrelated work.\n4. **Only dispatch independent steps.** Sequential dependencies stay with the parent.\n\n---\n\n**No gate.** Dispatch is a judgment call based on task size and parallelism opportunity.\n","# Subagent Dispatch\n\n**Principle:** The parent agent has full context (rules, memory, conventions). It\ncreates the plan. Subagents execute plan steps — they don't rediscover the project.\n\n---\n\n## When to Dispatch\n\n| Signal | Why |\n|---|---|\n| Plan has independent steps that can run in parallel | Concurrent execution saves time |\n| Context window is getting large | Fresh subagent avoids attention decay |\n| Research would pollute main context with raw file reads | Subagent returns a summary |\n| Code review needs a fresh perspective | Reviewer hasn't seen implementation reasoning |\n\n**Do NOT dispatch when:** steps have sequential dependencies, or the task is simpler\nto do directly than to explain.\n\n---\n\n## How to Dispatch\n\nThe parent creates the plan first (via task-planner or workflow Phase 1), then\nidentifies which steps can run in parallel.\n\n### Implementation dispatch\nPass the specific plan steps. The parent already checked rules, conventions, and\nsacred behaviors when creating the plan — the subagent doesn't need to re-read them.\n\n```\n\"Execute steps 3-5 from the plan:\n - Step 3: [exact instruction]\n - Step 4: [exact instruction]\n - Step 5: [exact instruction]\n Files to modify: [list].\n Do not commit or push.\"\n```\n\n### Review dispatch\nThe reviewer needs to know what to check against. Include references to constraints.\n\n```\n\"Review the changes made to [files].\n Read .memory\u002FSACRED.md — verify no sacred behaviors were modified.\n Read .memory\u002FCONVENTIONS.md — verify code follows project conventions.\n Check against the plan: [paste or summarize the plan].\n Report: issues found, verdict (approve\u002Frequest changes).\"\n```\n\n### Research dispatch\nPass the question. No rules needed.\n\n```\n\"How does [module X] connect to [module Y]? Read MAP.md for orientation.\n Report: answer, evidence (file:line), relevant files.\"\n```\n\n---\n\n## Validate Results\n\n- [ ] Read the subagent's output\n- [ ] Verify it meets the plan's success criteria\n- [ ] Check for side effects (unexpected file changes)\n- [ ] If validation fails, re-dispatch with corrected instructions\n\n---\n\n## Rules\n\n1. **Subagents do NOT run session-start.** They execute plan steps, not session init.\n2. **Subagents do NOT commit or push.** The parent handles git.\n3. **One task per subagent.** Don't batch unrelated work.\n4. **Only dispatch independent steps.** Sequential dependencies stay with the parent.\n\n---\n\n**No gate.** Dispatch is a judgment call based on task size and parallelism opportunity.\n",{"name":5,"license":15,"description":7},"main","d94c109a4fad51623dc6048315860e53b32673fb","core\u002Fskills\u002Fsubagent-dispatch",3,"https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel","https:\u002F\u002Fgithub.com\u002Fahmadulhoq","2026-11-07 00:53:22",{"source":8,"skillName":5,"commands":54,"prompt":64,"zipAvailable":65},[55,58,61],{"tool":56,"command":57},"npx","npx skills add ahmadulhoq\u002Fagentskel --skill subagent-dispatch",{"tool":59,"command":60},"bunx","bunx skills add ahmadulhoq\u002Fagentskel --skill subagent-dispatch",{"tool":62,"command":63},"pnpm","pnpm dlx skills add ahmadulhoq\u002Fagentskel --skill subagent-dispatch","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsubagent-dispatch\u002FSKILL.md",true,[],[68,87,111,135,159,172],{"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":15,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":41},"01M6GGSN2D57Y8BR50X08Y695H","setup-project","When setting up agentskel on a new project that does not yet have .memory\u002F or .agents\u002F. Use when the user says \"set up agentskel\", \"install agentskel\", or \"add agent memory to this project\".","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Fskills\u002Fsetup-project\u002FSKILL.md",[],[75,76,77,78,79,80,31,81,82,83,84,85,86],"setup project","project","setup","memory","add","up","install","add agent","user","new project","SET","agent memory",{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":41},"01M6GGSJ6HD2DFVZQTESJ81GYM","using-git-worktrees","When running long feature implementations, parallel branch work, or build\u002Ftest scripts that must not interfere with the user's active working directory. Creates isolated execution environments via git worktrees in sibling directories.","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Froles\u002Fdev\u002Fskills\u002Fusing-git-worktrees\u002FSKILL.md",[],[94,95,96,97,98,99,100,27,28,29,101,102,103,104,105,106,107,108,109,110],"git worktrees","worktrees","git","using git worktrees","Using git","test","directory","branch","directories","environments","with","via","long","scripts","feature","active","build",{"skillNo":112,"name":113,"nameZh":6,"description":114,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":115,"categories":116,"keywords":120,"publishedAt":41},"01M6GGSCFPK7MDSZH96XTNZR9B","test-engineer","When designing test strategy, analysing coverage, or writing tests outside the TDD cycle. Also use when working with CI pipelines or validating changes end-to-end.","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Froles\u002Fdev\u002Fskills\u002Ftest-engineer\u002FSKILL.md",[117],{"categoryNo":118,"categoryName":119,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",[121,122,99,123,124,125,126,127,128,31,129,104,130,131,132,133,134],"test engineer","engineer","tests","tdd cycle","cycle","end","test strategy","pipelines","tdd","strategy","ci","writing tests","writing","coverage",{"skillNo":136,"name":137,"nameZh":6,"description":138,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":139,"categories":140,"keywords":145,"publishedAt":41},"01M6GGS8YA9XZB6DWV22JC0YTB","test-driven-development","When implementing logic changes during develop-feature or implement-task. Use the Red-Green-Refactor cycle to write a failing test before any production code.","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Froles\u002Fdev\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[141,144],{"categoryNo":142,"categoryName":143,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":118,"categoryName":119,"taxonomyType":21,"taxonomyTypeDesc":22},[146,147,99,148,35,149,108,150,151,125,152,31,153,154,155,156,157,158],"development","test driven development","driven","red","logic","green","red-green-refactor","write","develop","refactor","Production","code","implement",{"skillNo":160,"name":161,"nameZh":6,"description":162,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":163,"categories":164,"keywords":165,"publishedAt":41},"01M6GGS61D41ZJG4D8XAS97H62","task-planner","When decomposing a feature or spec into implementable subtasks, estimating scope, or orchestrating multi-step work across subagents.","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Froles\u002Fdev\u002Fskills\u002Ftask-planner\u002FSKILL.md",[],[166,35,167,108,168,169,29,170,38,171],"task planner","planner","across","subagents","spec","multi",{"skillNo":173,"name":174,"nameZh":6,"description":175,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":176,"categories":177,"keywords":178,"publishedAt":41},"01M6GGS30NJW4Q7QN4KVT0CV08","systematic-debugger","When applying debugging techniques during any investigation — root cause tracing, bisect, defense-in-depth. Enforces hypothesis-first analysis and prohibits guess-and-check changes. Use standalone or inside the debug-issue workflow.","https:\u002F\u002Fgithub.com\u002Fahmadulhoq\u002Fagentskel\u002Fblob\u002Fmain\u002Froles\u002Fdev\u002Fskills\u002Fsystematic-debugger\u002FSKILL.md",[],[179,180,181,182,183,184,185,186,31,187,188,189,190,191,192,193,194,195,196,197],"systematic debugger","systematic","debugger","check","standalone","root","issue workflow","issue","bisect","investigation","analysis","first","defense","hypothesis","defense-in-depth","tracing","debugging","debug","workflow"]