Inspiration
When an appliance breaks, the evidence a technician needs (receipt, manual, model label photo, past service notes) is scattered across folders. People resend photos and guess model numbers, and old records often contradict each other.
What it does
Repair Evidence Dossier is a UGREEN AI NAS Skill that turns files the user explicitly authorizes into a source-traceable repair handoff for one appliance: identity, the symptom in the user's words, a dated timeline, a fact/value/status/source table, contradictions, missing evidence and questions for the technician. It organizes evidence; it does not diagnose faults or decide warranty questions.
How we built it
A Markdown-only Skill built with uliya-skill-kit 0.7.3. It delegates every NAS step to the official experts: file-agent (inventory), document-agent (receipts, manuals, notes), photo-agent (label OCR) and knowledge-agent (optional recording transcription, and save-text only after explicit approval). It passes the kit's static checker with no errors or warnings.
Challenges we ran into
Keeping every claim traceable, never guessing unreadable serials, keeping conflicting records side by side, and handling failed, cancelled and waiting states honestly.
What we learned
Strict delegate-only Skills need clear failure and approval handling at every step. Setelah itu isi Built with lalu tekan Save & continue.
What's next for Repair Evidence Dossier
Validation on a real UGREEN NAS if selected as a finalist. AI assistance: drafted with Codex and Grok Bot, reviewed by the entrant.
Built With
- codex
- markdownugreen-ai-nas
- ugreen-ai-nas
- uliya-skill-kit
