workflow-to-skill

To convert completed project workflows into reusable agent skills, provide the workflow slug to process JOURNEY.md, ADRs, and code files. This generates a draft SKILL.md that captures the actual sequence of steps, decisions, and known failures for future autonomous execution.

Documentation drift occurs when project records diverge from actual implementation. Capturing ADRs and JOURNEY.md files prevents drift by extracting architectural decisions and prerequisites into an auditable SKILL.md, preserving real-world learnings for future agent reuse.

Extracting prerequisites and safety constraints from ADRs involves analyzing completed architecture decision records to identify dependencies and prevent misconfigurations. The workflow transformation process identifies these constraints to ensure reliable automation and generate safety rules.

Yes, workflow conversion reconstructs and validates workflow provenance from JOURNEY.md, CLAUDE.md, and ADRs. These files are required to ensure the workflow is complete and to accurately capture the sequence of steps, decisions, and known failures for the skill artifact.

The best way to document known failures for autonomous agent reuse is to transform completed workflows into a SKILL.md artifact. This draft document explicitly captures known failures, safety rules, and validated steps from JOURNEY.md and ADRs to guide future automated executions.

Yes, you can review and approve the generated SKILL.md before finalization. The workflow conversion process supports a review workflow, allowing you to validate the drafted skill artifact, check extracted prerequisites, and confirm safety constraints before writing the final document.