What problem does it solve?
Digitized physical photo archives require a stable, machine-readable map linking original capture files to their published names, associated physical archive items, and rotation metadata, but manually creating and maintaining this mapping is error-prone and inconsistent across team members and sessions.
Core Features & Use Cases
- Human-authoritative input parsing: Accepts human-authored numbering documents in markdown pipe table or CSV format, with case-insensitive column matching so document structure can vary without breaking the workflow.
- Automatic metadata derivation: Derives the multi-item flag directly from stable output names (e.g., names ending in -multi set multi_item to true) and normalizes free-text rotation hints to canonical clockwise degree values, eliminating manual mapping errors.
- Strict read-only safety: Never modifies, opens, or processes the bytes of source capture files, only reading directory listings to detect missing captures, ensuring original archive assets remain untouched.
- Use Case: An archivist working with a collection of phone-captured scanned photos uses this skill to turn their handwritten numbering table into a consistent, version-controllable manifest that downstream image processing and publication tools can consume reliably.
Quick Start
Use the project-photos skill to generate a deterministic provenance manifest for your scanned photo archive using your human-authored numbering document and directory of original capture files.
