What problem does it solve? AI coding agents often produce unverified, bloated code with no engineering rigor. This Skill enforces a disciplined workflow where every task routes through a matched playbook, named engineering principles, delegated subagents, and runtime verification before any PR ships. ## Core Features & Use Cases - Playbook routing: Matches tasks (bug fix, feature, refactor, perf, investigation, babysit, shipping, autonomous runs) to step-by-step playbooks with mandatory verification gates. - Principle-driven decisions: Applies named principles like Laziness Protocol, Model the Domain, and Prove It Works, requiring each cited principle to trace to a real decision. - Subagent orchestration: Delegates code to model-tiered subagents (haiku, sonnet, fable) with swarm verification, PR size budgets, and review-bot triage. - Use Case: Ask the agent to fix a reported bug; it reproduces the defect on the real surface, binary-searches the root cause, delegates the fix, verifies the repro passes, and opens a right-sized PR. ## Quick Start Ask the agent to work in poteto mode on your next bug fix or feature, for example: use poteto-mode to investigate and fix the login timeout bug.