Claim-Ready

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for Claim-Ready

Inspiration

Insurers tell every household to keep a home inventory. Almost nobody does, because it means photographing every room and digging up receipts. After a burglary, fire or flood, people end up listing belongings from memory and can't prove what they paid. Yet the evidence usually already exists: phone photo backups and emailed or scanned receipts sit on the family NAS.

What it does

Claim-Ready is a Uliya Skill with three modes:

• Build or refresh the inventory. It finds household valuables in the photos you authorize, matches each one to a receipt, warranty or appraisal, and marks every item Proven, Photo only or Receipt only. It also gives a Claim-Ready score: the share of documented value that is fully proven, plus the three missing proofs that would raise it most.
• Coverage check. It reads your home or renters policy (limits, sub-limits, deductible, valuation basis) and flags gaps such as "jewelry exceeds the theft sub-limit". Each answer cites the policy section.
• Claim packet. After a loss ("my laptop was stolen last night"), it copies the proofs into one claim folder and re-verifies each file's SHA-256 fingerprint against the one recorded at inventory time. It quotes the deductible and reporting deadline and gives a clearly labeled estimate. Camera events from the loss window are added only if you opt in.

Every write (inventory file, album, tags, claim folder) waits for your approval. Nothing is deleted, moved or overwritten, and nothing leaves the NAS.

How we built it

We used the official UGOS Skill Creation Tool (uliya-skill-kit 0.7.3, capability catalog 2.0.0). The Skill is pure orchestration Markdown. The parent only delegates self-contained instructions to the registered experts (file-agent, photo-agent, document-agent, knowledge-agent, and security-monitor-agent as an option), and each step passes on only the real result of the step before it. Item matching follows explicit rules: serial or model first, then item type plus purchase date, then a match the user confirms. Prices are never estimated. Six linked reference files cover the inventory format, the coverage rules, the claim workflow, permissions and privacy, the validation plan, and disclosures. The package passes the kit's static checker with 0 errors and 0 warnings.

Challenges

• Writing a workflow that never invents a value: an unknown price stays "unknown" and is counted separately.
• Keeping every step inside the published capability catalog.
• Making refresh incremental (tags plus the date of the last inventory) so that repeat use is fast.

Accomplishments

• A single number people can act on (the Claim-Ready score).
• Tamper-evident proofs, re-verified at claim time.
• A privacy design suited to the most sensitive data in a home: faces are excluded, card numbers are masked, the local model is recommended, and a cloud model is used only with consent.

What we learned

On an AI NAS the strongest Skills combine several data types the user already owns. Photos and receipts are each ordinary on their own; joined, they become evidence.

What's next

If we're selected as a finalist, we'll run our synthetic test set on the iDX6011 Pro: two core cases (build, then refresh plus claim) and one edge case (empty sources, unrelated policy, no cameras). Then we'll tune the instruction wording to the real expert responses and record the demo.

Disclosures

• Third-party components and licenses: none. The package contains only original Markdown and uses no external services, APIs or datasets.
• AI tools: Grok Bot (AI assistant), used for rules and catalog research, drafting help and running the static checker. The UGOS Skill Creation Tool was used for the capability catalog and the static check.
• Test data is fully synthetic.

Built With

• ai
• markdown
• nas
• skill.md
• ugos
• ugreen
• uliya
• uliya-skill-kit