What problem does it solve? Lifecycle skills like create-spike, create-plan, and execute-plan need a resolved project context before they can act, but determining which project the user means from the working directory or a vague prompt is error-prone. This Skill centralizes that resolution, loading canonical project files or scaffolding a new project skeleton, so calling skills never guess. ## Core Features & Use Cases - Project resolution: Detects the target project from the current working directory (via git root and the docs/xNNN-name convention) or from an explicit name in the prompt, with exact and descriptive matching. - Context loading or scaffolding: Reads README.md, tasks.md, decisions/, and notes/ for existing projects, or invokes the yolo CLI to scaffold a new project skeleton with canonical structure. - Structured metadata and bounded refusal: Returns path, name, status, open spikes/plans, and decision/note counts to the calling skill, and refuses with a stable error shape when no project is determinable. - Use Case: A user runs /create-plan while inside docs/x003-auth/; launch-project resolves the project, loads its charter and open decisions, and hands metadata to create-plan so planning starts immediately. ## Quick Start Ask the agent to launch the project named x000-yolo-project so its context is loaded before running any lifecycle skill.
