What problem does it solve? Turning an approved OpenSpec change proposal into working code is error-prone: tasks get skipped, cited identifiers drift on main, and progress tracking is manual. This Skill orchestrates the implementation loop so each task in the change is executed, checked off, and verified against the current codebase. ## Core Features & Use Cases - Change selection and status parsing: Resolves the target change from context or openspec list --json, then reads schema, progress, and task state via openspec status and openspec instructions apply. - Preflight citation drift check: Re-greps identifiers cited in proposal, design, and tasks files against main to detect renamed classes, migration slots, or ADR numbers before writing code. - Task-by-task implementation loop: Works through pending tasks, marks - [ ] to - [x] after each completion, and pauses with structured options on ambiguity, errors, or design issues. - Use Case: A developer has an approved OpenSpec change with 7 tasks. They invoke the Skill, which writes the apply-session marker, verifies cited identifiers still exist on main, implements each task, and reports 7/7 completion ready for archive. ## Quick Start Ask the AI to start implementing the tasks from your OpenSpec change, optionally naming the change to apply.
