Portrait Consent Guardian

Inspiration

Commercial photographers, small creative studios, and schools store thousands of portraits on their NAS alongside signed model or parental release PDFs and field recordings. Before publishing a campaign or gallery, manually checking whether every visible face—including background bystanders—has an active release covering that specific channel and shoot date takes hours and risks legal takedowns.

What it does

portrait-consent-guardian is an orchestration Skill for UGREEN AI NAS (UGOS / Uliya) that verifies whether every identifiable person in a candidate NAS photo folder is cleared for a target publication channel and date by cross-referencing:

1. Local face clusters & EXIF timestamps (photo-agent via ug-photo figure-albums, album-detail, search-picture, meta, vision-vlm)
2. Signed release PDFs (document-agent via ug-document extract, qa, reconcile)
3. On-site voice memos (knowledge-agent via ug-knowledge audio-transcription, collect-evidence)
4. Scoped file enumeration & approval-gated tagging (file-agent via ug-filesystem find, ls and ug-filemgr tag-add, tag-add-file)

How we built it

We designed SKILL.md and two supporting references (references/consent-decision-matrix.md and references/compliance-and-disclosures.md) strictly following the Uliya Skill Development Standard v2.0.0 and validated the package against check-skill.mjs with static-pass (0 errors, 0 warnings).

Built With

• ai-agents
• markdown
• ugos
• uliya