What problem does it solve? Turning approved requirements into a concrete, codebase-grounded solution design is slow and error-prone when done ad hoc. This Skill automates the design stage of a planning workflow: it gathers codebase context, defines how the work will be verified, and writes a solution design into the task's plan.html document. ## Core Features & Use Cases - Parallel Context Gathering: Spawns two subagents (code and docs) to ground the design in real codebase patterns, then writes a context.md file. - Verification Tab Generation: Designs a test strategy with Given/When/Then acceptance criteria inserted as a Verification tab in plan.html. - Solution Tab with Tradeoffs: Explores multiple approaches via parallel subagents when the task is ambiguous, records rejected alternatives, and writes the chosen design as a Solution tab. - Interactive and Autonomous Modes: Asks the user to resolve open questions when interactive, or records decisions as gated pd-question elements with recommended answers when running unattended. - Use Case: After requirements for a new API endpoint are approved, run this Skill to produce context.md plus Verification and Solution tabs in plan.html, ready for review before the planning stage. ## Quick Start Ask the agent to write a solution for the current task, for example: "Design the solution for the approved requirements in the latest task folder."
