jbaehova

Orchestrate Claude Code subagents for explicit fan-out work across discovery, implementation, testing, validation, review, and synthesis. Use only when the user invokes /fan-out or asks Claude to fan out, spawn subagents, delegate work in parallel, run separate agents per question/subsystem/file group/hypothesis, or synthesize parallel work. Do not use for ordinary complex tasks unless parallel delegation is explicitly requested.

Use when the user explicitly invokes $fan-out or asks Codex to fan out, spawn subagents, delegate work in parallel, use explorer/worker agents, or assign one agent per question, subsystem, file group, or hypothesis. Orchestrate built-in Codex subagents across discovery, implementation, testing, validation, review, and integration for non-trivial coding, investigation, or delivery tasks, then synthesize one final response. Do not use for ordinary complex tasks unless parallel delegation is explicit.

Orchestrate Cursor subagents for explicit fan-out work across discovery, implementation, testing, validation, browser checks, command analysis, review, and synthesis. Use only when the user invokes /fan-out or asks Cursor to fan out, spawn subagents, delegate work in parallel, run separate agents per question/subsystem/file group/hypothesis, or synthesize parallel work. Do not use for ordinary complex tasks unless parallel delegation is explicitly requested.

Delegate host coding-agent subtasks to read-by-default Pi CLI model workers with optional direct-write permission, provider/model/thinking aliases, routes, parallel execution, write capture, and optional session continuity. Use when Codex, Claude Code, Cursor, Hermes, or another Agent Skills-compatible host should consult specialist models for writing, design critique, code review, planning, alternative patches, direct edits, or model comparison.

Analyze PDF files by converting them into agent-readable text, OCR text, page render images, and extracted embedded images without polluting the working directory. Use when an agent needs to read, summarize, inspect, compare, quote, or reason about PDFs, especially when the foundation model cannot ingest PDFs directly or when tables, figures, scans, screenshots, or layout matter.