Inspiration

Family photo libraries can contain years of memories that are difficult to organize into a story. Story Reel proposes a NAS-local workflow that uses photo evidence and explicit approval before creating an album.

What it does

The skill asks for a period, event, people, or place and an output choice. It delegates discovery and clustering to the photo expert, then asks the document expert for a factual two-to-five-chapter narrative using only returned evidence. The user reviews the draft before the photo expert creates the album and the knowledge expert saves the story document. Missing evidence is marked unknown.

How it is built

Story Reel is a Markdown orchestration skill following the Uliya Skill Development Standard. It uses ug-photo, ug-document, and ug-knowledge through registered experts. The parent skill declares its capabilities, handles empty and partial results and delegation states, and stops dependent steps when inputs are missing.

Challenges and lessons

The main design challenge was preserving evidence across expert handoffs without inventing captions or treating a partially completed operation as success. The package includes explicit recovery states and approval checkpoints.

Current status

This Round 1 package is statically validated with the official checker: zero errors and zero warnings. It has not yet run on the assigned UGREEN AI NAS. The proposal PDF specifies two core tests and an edge case for finalist testing.

What is next

If selected, validate the expert round trips and real-library behavior on the assigned iDX6011 Pro during the testing day, capture evidence, and record the Round 2 demonstration.