What problem does it solve? AI coding agents often produce verbose prose, unverified claims, and ad-hoc changes without a consistent process. This Skill imposes a rigorous operating style: named engineering principles, task-matched playbooks, deliberate subagent delegation, and evidence-backed verification before anything ships. ## Core Features & Use Cases - Task-matched playbooks: Routes work to Investigation, Bug fix, Perf issue, Feature, Refactoring, Prototype, or Shipping playbooks, each with verbatim step lists and reply formats. - Named principles library: Applies core, architecture, verification, delegation, and meta principles (e.g., Laziness Protocol, Model the Domain, Prove It Works) with explicit citation of each principle that shaped a decision. - Subagent discipline: Prescribes background subagents, model roles for adversarial review, and owner-side diff review so delegated work is verified, not trusted. - Use Case: A developer asks the agent to fix a reported bug. The agent reproduces it via a control skill, binary-searches the root cause with runtime evidence, delegates the fix, verifies on the same surface, and opens a PR with a structured description. ## Quick Start Invoke /poteto-mode and ask the agent to fix a bug or build a feature so it follows the matching playbook end to end.
