What it does

unphotographed-years inventories a family photo archive year by year, breaks the three thinnest years down month by month, names the person appearing least, and saves one cited report.

Most family archives are not empty. They are uneven. Ten years of photos can hold three rich years and four nearly blank ones, and the blank years are usually the ones nobody can recall.

Try it

  • "Which years are missing from our family photos?"
  • "What did our photo archive fail to capture?"
  • "Check the coverage of my photos from 2012 to 2026."
  • "Who is absent from our family photos?"

How it works

  1. Year coverage — ug-photo / photo-agent returns a count per year across the covered span, and every zero-photo year is listed explicitly.
  2. Month coverage — the three thinnest years are broken into all twelve months, zeros marked.
  3. People coverage — ug-photo figure albums give per-person counts, unnamed faces as one group, plus multi-year interruptions.
  4. The report — ug-knowledge / knowledge-agent writes one plain text report into knowledge storage after you confirm the destination file name.

Each step consumes only the real result of the step before it. Nothing is estimated, interpolated, or invented to fill a gap.

Boundaries

  • Read-only by default. The only write is the report, and it asks you to confirm the destination file name first.
  • No deletions. It never deletes, moves, renames, or re-tags anything.
  • Local only. On-device photo library and knowledge storage. No external service.
  • Photos only. No surveillance, video, or documents.

Status

Round 1: SKILL.md with frontmatter, scope, inputs, capability evidence, a four-step orchestration, a completion/failure table, empty- and partial-result rules, and three linked references. The bundled static checker reports zero errors. Nothing has run on hardware yet — NAS validation will cover count reliability, the save-text destination, and latency.

AI tools used

An AI coding assistant drafted and revised the Markdown orchestration and ran the bundled static checker during authoring. No AI component runs inside the Skill; at run time the registered on-device NAS experts do all the work.

Built With

  • markdown
  • ug-knowledge
  • ug-photo
  • ugos
  • ugreen-ai-nas
  • ugreen-skill-development-standard
Share this project:

Updates

Submission history