This converts your PRD markdown into a structured epic and child beads that ralph-tui can execute autonomously. It extracts quality gates from your PRD, splits user stories into right-sized tasks (completable in one agent context window), sets up dependencies so schema changes happen before backend before UI, and outputs beads in .beads/beads.jsonl format. The heredoc syntax handling is solid, preventing shell interpretation issues with backticks and variables in descriptions. The one-story-per-iteration sizing guidance is critical since ralph-tui spawns fresh agents with no memory between runs. If you're planning to feed PRDs to ralph and want proper task decomposition with verification gates baked in, this handles the busywork of creating that structure.
npx -y skills add subsy/ralph-tui --skill ralph-tui-create-beads --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills