What problem does it solve? Turning a research plan, review finding, or conversation into work that an AI coding agent can execute safely is error-prone: scope creeps, acceptance criteria stay vague, and dispatch happens on unverified assumptions. This Skill converts plans into bounded, vertical task briefs for Claude Code or Codex with machine-checkable acceptance criteria. ## Core Features & Use Cases - Structured Brief Template: Produces task briefs covering execution context, assurance lanes, upstream contracts, inputs/outputs by path, validation commands, provenance, runtime constraints, and stop conditions. - Dispatch Safety Rules: Enforces hard scope stops, blocking user-decision gates, a no-synthetic-results rule, and verification steps for any inferred external dependency. - Boundary Routing: Routes outcome-contingent work to pre-registration and unresolved concepts to domain modeling instead of dispatching prematurely. - Use Case: A research manager has a plan note for a statistical analysis pipeline. Use this Skill to produce a task brief naming the capability gap, required contracts, validation commands, and stop conditions before handing it to an implementing agent. ## Quick Start Convert this research plan into an agent-ready task brief with acceptance criteria, assurance lanes, and stop conditions.
