Menu
Route delegated work through the current host's supported sub-agent mechanism. In Pi, use the installed pi-subagents package. Use whenever the user asks for a subagent, Pi agent, Claude or Codex agent, or independent multi-agent work.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Launch both thermo-nuclear review subagents in parallel. Use for thermos, double thermo review, or combined bug/security and code-quality branch audits.
Run the CodeRabbit CLI for AI-powered review. Use only when the user explicitly asks for CodeRabbit or invokes coderabbit-review. Never trigger for a generic code review, PR review, quality check, security audit, or autonomously.
Use when reviewing implementation plans, migration plans, refactoring proposals, feature plans, generated implementation trackers, or committable phase plans before execution; especially when asked whether a plan is ready, concrete enough, testable, safe, or dependency-aware.
Use when creating, refining, or reviewing implementation plans, migration plans, refactoring proposals, feature plans, implementation trackers, or committable phase plans; especially when turning a rough plan into executable phases or checking whether a plan is concrete, testable, safe, dependency-aware, and suitable for parallel work.
Use this skill whenever the user gives a YouTube URL and asks to summarize, digest, analyze, extract transcript/context, decide whether a video is worth watching, or understand what a video is about. Prefer the summarize CLI for YouTube/video context instead of web search. Use Codex as the default summarize backend when possible, and use the local Whisper large-v3 model for transcription fallback. Do not use this for ordinary web pages; the user prefers Exa for webpage research.
Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth. Use for a thermo-nuclear code quality review, thermonuclear review, deep code quality audit, or especially harsh maintainability review.
| name | subagents |
| description | Route delegated work through the current host's supported sub-agent mechanism. In Pi, use the installed pi-subagents package. Use whenever the user asks for a subagent, Pi agent, Claude or Codex agent, or independent multi-agent work. |
When the subagent tool is available, use it and follow the bundled
pi-subagents skill for single, parallel, chained, background, and
forked-context delegation.
Treat explicit Claude or Codex wording as a model/provider preference inside Pi, not as a native CLI harness. Use an exact available model when the request identifies one; otherwise ask rather than guessing.
Do not use legacy start_subagent, stop_subagent, or reply_subagent tools.
Use only the current host's documented native delegation mechanism. Do not claim Pi package behavior or silently substitute a different host.
