Autonomous multi-agent pipeline that lifts any coding agent to senior-engineer discipline on complex or high-stakes work. Use when the user wants thorough, self-reviewed, production-grade output — implementing features, fixing hard bugs, refactoring, designing architectures or APIs, planning migrations, or deep research — or asks for titan, maximum rigor, or full-pipeline treatment. Runs guard → route → decompose → recall → ladder → tree-search → execute → reflexion → verify → council → gate end to end without babysitting, using the host agent's own subagents and reasoning. No API key or external service required.
