What problem does it solve? It removes the guesswork of turning an OpenSpec change plan into working code by walking through each task in order, tracking progress, and pausing safely when blocked. ## Core Features & Use Cases - Change selection and status checks: Resolves the target change from user input or context, then reads schema, progress, and task state via openspec status and openspec instructions apply. - Guided task implementation loop: Reads the context files (proposal, specs, design, tasks), implements each pending task, and marks checkboxes complete as it goes. - Blocked and completion handling: Detects blocked or all-done states, suggests the next artifact or archive step, and reports progress after every session. - Use Case: A developer with a spec-driven change containing 7 tasks asks the agent to continue implementation; the skill reads the tasks file, implements tasks 4 through 7, checks them off, and suggests archiving the change. ## Quick Start Ask the agent to implement the tasks for your OpenSpec change, for example by saying "apply the add-auth change", and it will select the change, read its context files, and work through the remaining tasks.
