What problem does it solve? AI coding agents often jump straight into implementation without a documented rationale, scope boundary, or rollback strategy. This Skill enforces a spec-driven development discipline by generating a structured proposal.md artifact that captures intent, scope, capabilities, risks, and success criteria before any code is written. ## Core Features & Use Cases - Structured Proposal Generation: Produces a proposal.md with intent, in/out scope, capabilities contract, approach, affected areas, risks, rollback plan, and success criteria. - Multi-Mode Persistence: Supports engram, openspec, hybrid, and none artifact store modes, writing to Engram topics, openspec/changes directories, or returning inline results. - Interactive Proposal Shaping: Runs a product question round covering business rules, edge cases, and tradeoffs before finalizing the proposal. - Use Case: An orchestrator agent receives a request to add dark mode; it delegates to this Skill, which produces a concise proposal under 450 words defining the new capability contract that downstream spec and design agents consume. ## Quick Start Delegate a change proposal to the sdd-propose sub-agent with the change name and exploration analysis to generate a structured proposal.md.
