What problem does it solve? Confident answers are not always correct ones, and long sessions quietly turn assumptions into facts. This Skill catches wrong directions in non-trivial decisions while course-correction is still cheap, instead of discovering them in production. ## Core Features & Use Cases - Five-step doubt cycle: CLAIM the decision, EXTRACT the smallest reviewable artifact plus contract, DOUBT it with a fresh-context adversarial reviewer, RECONCILE findings by precedence, and STOP at a bounded condition (trivial findings, 3 cycles, or user override). - Adversarial reviewer prompting: Passes only the artifact and contract to a fresh-context reviewer with an issues-only prompt, never the author's conclusion, avoiding validation bias. - Cross-model escalation: In interactive sessions, explicitly offers a second opinion via external CLIs such as Gemini or Codex with read-only sandboxing and per-invocation user authorization. - Use Case: Before committing a new caching layer claimed to be thread-safe, run the doubt cycle so an isolated reviewer hunts for race conditions, unstated assumptions, and contract violations before the code ships. ## Quick Start Ask the agent to apply doubt-driven development to the architectural decision or non-trivial code change you are about to commit.
