What problem does it solve? Turning a design spec or approved requirements into an executable implementation plan is error-prone: plans often contain vague steps, missing test commands, or placeholders that force the implementing agent to guess. This Skill produces concrete, slice-ordered plans that a fresh agent can execute step by step without conversation history. ## Core Features & Use Cases - Slice-based decomposition: Orders work as vertical, end-to-end increments starting with a tracer-bullet slice, each with exact file paths, type signatures, test commands, and acceptance criteria. - Placeholder prevention and review: Forbids TBD-style steps and dispatches an adversarial reviewer subagent (using the included prompt template) to verify spec coverage, name consistency, and testability before handoff. - Flexible output formats: Emits Markdown by default or styled HTML via the bundled template, with optional design-system overrides and configurable plan paths. - Use Case: After approving a spec for a new API feature, invoke this Skill to produce a plan document with checkbox TDD steps per slice, then hand it to a subagent-driven execution workflow. ## Quick Start Ask the agent to write an implementation plan for the approved spec at .workbench/specs/my-feature.md using the writing-plans skill.
