Inspiration
Legal teams work with evolving case files: claims, responses, court orders, and supporting documents. Each new filing requires understanding what changed without losing the history of the case. My experience with legal workflows inspired NYVARA Case Tracker: a way to organize that information while keeping professionals in control.
What it does
NYVARA Case Tracker is an AI skill designed for UGREEN NAS. It reviews an authorized case folder and prepares a case summary, chronology, and update report with references to the source documents.
The workflow distinguishes a party’s allegation from a court decision, flags missing information, and identifies documents belonging to another case. When a previous report is available, it compares the current evidence against that baseline.
How I built it
I used the Uliya Skill Kit to define the workflow in SKILL.md, supported by a report contract and a validation plan.
The design delegates file discovery, document extraction, and approved report saving to the supported Uliya experts. Original documents remain unchanged, and users review the report before approving its storage.
The package passed the kit’s static checker with zero errors and zero warnings. This checks the package structure; it does not establish runtime reliability.
Test scenario
I prepared a fully fictional legal case involving a promissory note. The initial documents establish the information available for the first summary.
Later documents introduce an alleged partial payment and a court order requesting supporting evidence. The expected behavior is to report the allegation without treating the payment as proven or automatically reducing the claimed amount.
Additional fixtures test an unrelated case, a duplicate document, and an unreadable file.
Challenges and lessons learned
The main challenge is preserving the meaning and authority of each document. A filing date is not necessarily a notification date, and a claimed payment is not a confirmed payment.
I learned to define explicit evidence requirements, uncertainty handling, and approval steps before relying on generated summaries. Local storage alone also does not guarantee local AI processing: the available model and any external data flow must be verified.
Current status and next steps
The skill package, synthetic documents, and validation plan are prepared. Execution on UGREEN NAS remains untested.
The next step is to evaluate extraction accuracy, source references, update comparisons, and approved saving in the supported NAS environment. NYVARA supports professional document review; it does not provide legal advice, calculate procedural deadlines, or automatically monitor courts.## 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 NYVARA Case Tracker
Built With
- kitugreen
- markdown
- nas
- skill
- uliya
