What problem does it solve?
Re-running setup can inadvertently overwrite or inconsistently handle previously generated assets, causing nondeterministic builds, lost manual edits, and confusion about provenance; this Skill provides a deterministic protocol to detect, classify, and safely reconcile those assets.
Core Features & Use Cases
- Provenance Classification: Detects generated vs manual assets using YAML frontmatter, meta blocks, and provenance markers across agents, skills, workflow.yaml, and CLAUDE.md variants.
- Disposition Actions & Preflight: Presents a preflight report with
[created], [updated], [skipped], and [migrated] dispositions and lets users approve selective merges or regenerations.
- Safe Merge & Optimization: Applies selective deep-merge rules for workflow.yaml, supports source-hash generation-time skips, and preserves user-owned keys while updating template-owned keys.
- Orphan Detection & Reporting: Flags context-file orphans and migration artifacts for user review without auto-deletion.
- Use Case: Use when re-running structured-workflows:setup on a repo to avoid overwriting user-authored CLAUDE.md content while updating generated skills and agents.
Quick Start
Run asset-reconciliation during a structured-workflows setup re-run to detect, classify, and propose dispositions for existing assets and present a preflight report.