What problem does it solve? Multi-step automation recipes often fail because they lack clear stop conditions, loop exits, branch logic, and approval gates, causing workflows to wander, loop forever, or fire irreversible actions without asking. This Skill provides the design discipline and structure needed to write reliable Claude Code workflow definitions. ## Core Features & Use Cases - Layer selection guidance: Decide whether a task warrants a workflow versus a command, agent, or skill before writing anything. - Structured workflow anatomy: Enforces purpose lines, inputs/outputs, ordered steps with asset references and hand-offs, control flow, STOP conditions, and an Invokes manifest. - Safety gates and anti-patterns: Requires approval gates on irreversible actions and flags common failures like unbounded loops, restated mandates, and chat-only state. - Use Case: You want to automate a release pipeline that verifies a build, tags a version, and opens a PR. Use this Skill to scaffold the workflow file with explicit success, bail, and gate conditions for each step. ## Quick Start Ask the assistant to author a workflow that orchestrates your chosen commands and agents into a multi-step pipeline with explicit stop conditions.
