What problem does it solve? Writing implementation plans without understanding the actual codebase leads to vague, unactionable specs. This Skill guides an AI agent through a structured, interactive process of reading tickets, researching code, and producing phased implementation plans with verifiable success criteria. ## Core Features & Use Cases - Interactive Plan Development: Gathers requirements, asks focused questions, and iterates on plan structure with the user before writing details. - Parallel Codebase Research: Spawns research sub-agents (codebase-locator, codebase-analyzer, thoughts-locator) to find relevant files, patterns, and prior decisions with file:line references. - Structured Plan Output: Writes plans to thoughts/shared/plans/ using a dated, ticket-linked filename convention with phases, automated and manual success criteria, and explicit out-of-scope items. - Use Case: A developer receives ticket ENG-1478 for parent-child task tracking. The Skill reads the ticket, researches the daemon codebase, proposes design options, and produces a phased plan with make test verification steps. ## Quick Start Ask the agent to create an implementation plan for a ticket file, for example: create a plan for thoughts/allison/tickets/eng_1234.md.
