What problem does it solve? Confident answers are not always correct ones, and long coding sessions quietly turn assumptions into accepted facts. This Skill catches wrong architectural and implementation decisions early — while course-correction is still cheap — by forcing every non-trivial decision through an adversarial fresh-context review before it is committed. ## Core Features & Use Cases - Five-step doubt cycle: CLAIM the decision, EXTRACT the smallest reviewable artifact plus its contract, DOUBT it with an adversarial reviewer, RECONCILE findings by precedence, and STOP within a bounded three-cycle loop. - Adversarial reviewer protocol: Spawns an isolated-context reviewer (or a shipped persona like code-reviewer or security-auditor) with an issues-only prompt, passing the artifact and contract but never your conclusion. - Cross-model escalation: Offers an optional second opinion via external CLIs such as Gemini or Codex with read-only sandboxing, stdin piping, and explicit per-invocation user authorization. - Use Case: Before committing a new caching layer you believe is thread-safe, run the doubt cycle — the reviewer probes unstated assumptions, race conditions, and contract violations, and you classify each finding as actionable, trade-off, or noise before merging. ## Quick Start Ask the assistant to apply doubt-driven development to adversarially review the caching layer change before committing it.
