What problem does it solve? Turning a change request into an executable implementation plan usually means scattered chat notes, missed cross-repo dependencies, and invented verification commands. This Skill investigates the repository first, resolves blocking ambiguities with the user, and writes one complete, validated YAML plan artifact that another agent can execute without repeating discovery. ## Core Features & Use Cases - Investigation-first planning: Reads project docs, git history, contracts, and code before asking only decision-bearing questions, never asking for facts discoverable in the repo. - Canonical YAML plan artifact: Writes exactly one plans/active/<slug>.yaml file with goal, acceptance criteria, ordered repos and steps, seams, risks, rollback, and decisions, with no placeholders allowed. - Multi-repo merge ordering: Derives repository order from contract ownership so seam owners land before consumers, and validates the result with evo harness commands when available. - Use Case: A user asks to plan a feature spanning an API service and its client SDK. The Skill traces the contract seam, asks one round of clarifying questions about rollout and compatibility, then writes a validated plan with owner-before-consumer merge order and per-step verification commands. ## Quick Start Ask the agent to create an execution plan for your change, for example: create an execution plan for adding OAuth login across the auth service and web app.
