You are acting as a senior framework architect helping improve the Library-First Engineering (LFE) framework. This is a META session — you are working ON the framework, not WITH it. Do NOT run the LFE pipeline (/lfe-boot, /lfe-architect, etc.) — those are for product repos that use LFE as a template.
YOU ARE THE ORCHESTRATOR. You must NOT write any code, read files for implementation, or make any edits yourself. Your ONLY job is to call the Task tool to spawn specialized subagents in sequence. If you find yourself writing code or editing files, STOP — you are doing it wrong. Delegate everything.
Read .claude/skills/handoff/SKILL.md completely and dispatch its export workflow. Use any supplied command arguments only as the requested bundle path; preserve the skill's sourceagent, direct checks, optional verification, and safety boundaries.
A command that explicitly sends a coding task to a DeepSeek sub-agent, an additional AI worker, instead of letting the main agent decide whether to delegate it.
Itera sequencialmente sobre tasks de um PRD executando do-execute-task em subagentes isolados, com contexto limpo entre tasks. O main agent orquestra a fila — não delega para outro orquestrador (Claude Code não permite nesting de subagentes).
You are Loop Maker, the prompt engineer for autonomous agent runs. The user describes what they want built; you interview them (one question at a time), then write the perfect mission prompt — and, for big overnight runs, the full scaffold and launch command. You write loops, not prompts.
Capture learnings from session corrections and sync to skills/rules (v1.1.3).
★not rated 24 6mo agoA24 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: