Skip to content
Get packs

Subagent-Driven Development

Runs a multi-task plan by handing each task to a fresh sub-agent, with reviews after each step.

Source Download skill
Repository
jnMetaCode/superpowers-zh
Stars
8,444
Licence
MIT
Files
4 files

What it does

When an AI agent has an implementation plan with mostly independent tasks, this skill tells it to delegate each task to a fresh sub-agent with a clean context, then have separate reviewers check spec compliance and code quality after each task and once more across the whole branch. It keeps a progress ledger file so work can resume after the context is compacted, and it sets rules for choosing cheaper or stronger models per task and for continuing without stopping to ask the user.

Great for

  • Developers who have a written multi-step implementation plan and want it executed in one session with review checkpoints
  • Teams experimenting with multi-agent workflows who want isolated context per task
  • People comfortable letting an agent make routine decisions and record them in a ledger

Skip it if

  • Tightly coupled work where tasks depend heavily on each other, since the skill itself says to brainstorm or work manually in that case
  • Situations where you want to approve each decision, because the skill tells the agent to keep going and rule on conflicts itself, which can produce rework if it guesses wrong
  • Environments without sub-agent support, where the skill is explicitly not meant to be used

Try asking

  • “Execute the implementation plan in docs/plan.md using subagents, one per task, with reviews after each”
  • “Resume the plan I started yesterday from the progress ledger without redoing finished tasks”
  • “Run this multi-step refactor plan end to end and do a final whole-branch review at the end”
Step-by-step guide: how to use Subagent-Driven DevelopmentInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 4 original files, the MIT licence and an attribution note.

    subagent-driven-development.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip subagent-driven-development.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip subagent-driven-development.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Write a clear implementation plan with a spec before using it, since the agent resolves conflicts against the plan and spec and conflicts it rules on are logged rather than asked about
  • Review the progress ledger at .superpowers/sdd/ in your repo periodically, since the agent makes and records its own rulings and fixes
  • Work on a feature branch or worktree, as the skill says not to start on main or master without explicit permission

What's inside

  • implementer-prompt.md
  • re-review-prompt.md
  • SKILL.md
  • task-reviewer-prompt.md

Redistributed unmodified from jnMetaCode/superpowers-zh under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.