What problem does it solve? When an issue moves through an automated SDLC pipeline, someone must turn the spec and research findings into a concrete, ordered implementation plan. This Skill acts as the Plan stage agent in the Fabrik pipeline, producing a plan an implementer can follow mechanically without making design decisions. ## Core Features & Use Cases - Implementation Planning: Designs the approach, file organization, interfaces, and testing strategy based on the issue spec and Research stage output. - Task Checklist Generation: Produces an ordered, atomic, testable GitHub-checkbox task list (typically 5-15 tasks) with documentation updates integrated alongside code tasks. - Sub-issue Decomposition: Emits FABRIK_SPAWN_CHILD blocks so the engine can split multi-repo or parallelizable work into linked child issues with optional DEPENDS_ON ordering. - Pipeline Signaling: Uses FABRIK_STAGE_COMPLETE, FABRIK_NO_WORK_NEEDED, and FABRIK_BLOCKED_ON_INPUT markers to drive the engine's stage transitions. - Use Case: An issue requesting a new API endpoint reaches the Plan column; the agent reads .fabrik-context/issue.md and stage-Research.md, decides the approach, and posts a plan with a 10-task checklist that the Implement stage executes. ## Quick Start Read the spec and research files in .fabrik-context, then produce an implementation plan with a task checklist and emit FABRIK_STAGE_COMPLETE.
