Skip to content
ai101.tools
navigateopenescclose
Claude+372Whisper+228LangChain+168Codex+223NotebookLM+276DALL-E 3+192DeepL+249n8n+208Topaz Video AI+153LlamaIndex+161
Subagent-Driven Development logo
Skill

Subagent-Driven Development

Hands each plan task to a fresh subagent and gates it behind a mandatory review.

0
SaveVisit website ↗

Install

Ships scripts
/plugin install superpowers@claude-plugins-official
Triggers on

Used when a plan's tasks are independent enough to hand each one to its own subagent inside the current session.

Author
Jesse Vincent (obra)
License
MIT

Every task goes to an implementer that starts with clean context, so history from one task never leaks into the next and the coordinating agent keeps its own window free for orchestration. A finished task does not count as done until it clears a review on both spec compliance and code quality, with a bounded fix loop when findings come back. A running ledger records rulings and completed tasks so the process survives a context reset, and the whole branch gets one last review at the end.

Comments(0)

Sign in to comment

No comments yet — be the first.

Similar skills

Report this comment

Why are you reporting this?