{
  "id": "subagent-driven-development",
  "handle": "subagent-driven-development",
  "canonicalUrl": "https://bestskillsforclaude.com/en/skills/subagent-driven-development",
  "language": "en",
  "alternates": {
    "en": "https://bestskillsforclaude.com/en/skills/subagent-driven-development",
    "es": "https://bestskillsforclaude.com/es/skills/subagent-driven-development",
    "pt-BR": "https://bestskillsforclaude.com/pt-br/skills/subagent-driven-development",
    "ja": "https://bestskillsforclaude.com/ja/skills/subagent-driven-development"
  },
  "category": {
    "id": "planning-and-teamwork",
    "name": "Planning and teamwork"
  },
  "publisher": {
    "name": "Jesse Vincent (obra)",
    "url": "https://github.com/obra"
  },
  "source": "https://github.com/obra/superpowers",
  "documentation": null,
  "licenseAsStated": "MIT",
  "runtime": "Claude Code and other supported coding agents",
  "lastReviewed": "2026-08-05",
  "headline": "When does splitting work across agents actually help?",
  "summary": "When the pieces are genuinely independent, and almost never otherwise.",
  "verdict": "The subagent-driven-development skill structures work so that several agents can run concurrently without colliding. It pays off when the pieces are truly independent, such as separate modules or parallel investigations, and it produces merge pain and contradictory decisions when they are not.",
  "bestFor": "Wide, independent work: parallel research, per module changes, large mechanical sweeps.",
  "skipIf": "The pieces share state, interfaces or design decisions that have not been made yet.",
  "keyTakeaways": [
    "Independence is the precondition, not an optimisation detail.",
    "Splitting before planning creates parallel versions of the same misunderstanding.",
    "Merging is the real cost, and it is paid by you, not the agents.",
    "Read only fan out, such as research or review, is the safest and most reliable use."
  ],
  "quickStart": [
    "Install superpowers from obra/superpowers, published under MIT.",
    "Write the plan first, then split it. Splitting an unplanned task multiplies the confusion.",
    "Give each agent a boundary in files and a definition of done, and use isolation when they write to the same repository."
  ],
  "sections": [
    {
      "heading": "What it changes in practice",
      "paragraphs": [
        "Parallel agents are attractive for the obvious reason and dangerous for a less obvious one: they each make small design decisions, and nothing reconciles those decisions afterwards except you.",
        "The reliable pattern is fan out for reading and gathering, then a single writer for the decisions. Investigations, audits and reviews parallelise cleanly. Interface design does not."
      ]
    },
    {
      "heading": "Where it disappoints",
      "paragraphs": [
        "Shared files turn concurrency into conflict resolution, which is why worktree style isolation matters as soon as more than one agent writes.",
        "Coordination cost also grows faster than the speedup. Beyond a handful of concurrent pieces, the merge is the project."
      ]
    }
  ],
  "faq": [
    {
      "question": "When should I not parallelise?",
      "answer": "When the pieces share state or unmade design decisions. You will get several confident, incompatible answers."
    },
    {
      "question": "What is the safest use?",
      "answer": "Read only work such as research, audits and reviews, where nothing has to be merged."
    },
    {
      "question": "Do agents need isolated workspaces?",
      "answer": "Once more than one writes to the repository, yes. Otherwise you are managing conflicts by hand."
    }
  ],
  "sources": [
    {
      "label": "obra/superpowers on GitHub",
      "url": "https://github.com/obra/superpowers",
      "kind": "repository"
    },
    {
      "label": "Agent skills directory listing",
      "url": "https://skills.sh",
      "kind": "directory"
    }
  ],
  "machineReadable": {
    "markdown": "https://bestskillsforclaude.com/en/skills/subagent-driven-development/content.md",
    "json": "https://bestskillsforclaude.com/en/skills/subagent-driven-development/skill.json"
  },
  "disclosure": "Editorial desk review based on the publisher's own repository. Not a security audit, a benchmark or a guarantee of maintenance."
}