Loading Trouble Tracker…
Trouble Tracker
Admin-only. Sign in with your administrator Google account to continue.
Admin access required
The Trouble Tracker is restricted to administrators during this phase. Head back to the cockpit.
Trouble Tracker
Issues, features, parking-lot items, and Clinical Chats — all in one place with AI-augmented triage.
Every action logged to the chain-of-custody ledger.
Open issues
0
New + Triaged + In progress + Blocked
Done this week
0
Resolved last 7 days
Parking lot
0
Future consideration
Critical risk
0
AI-classified, needs review
Avg hours to close
3.2h
Trailing 30 days
Clinical Chats
0
Active threads with clinician input
ID
Priority
Title
Status
Type
Est.
Risk / Chats
ID
Priority
Title
Reason
Type
Est.
Risk / Chats
Parking Lot.
Distinct from "Won't fix" — these are items we deliberately deferred for consideration later.
Each carries a "Reason" field (defer-to-Phase-II, awaits-funding, awaits-IRB, blocked-on-vendor, etc.)
plus the AI analysis from when it was logged. Pulling something out of Parking Lot promotes it
back to Triaged status and re-runs the AI analysis with current codebase context.
Clinician-initiated threads not tied to a specific issue.
For practice-of-care conversations, "what works in PTSD population," questions from
the field. Tagged for searchability; available to clinicians at the partner role.
ID
Priority
Title
Closed
Type
Actual
Variance
AI estimation feedback loop.
Every closed issue compares estimated vs. actual hours. The "Variance" column drives a
rolling-average bias correction the AI applies to future estimates. This is what makes
the AI analysis improve over time — and a candidate distinguishing feature for Patent R+
extensions covering AI-augmented issue triage with accuracy-tracking feedback.
ID
Priority
Title
Status
Type
Est.
Risk / Chats
All items, every status.
A single sweep view across Issues, Parking Lot, and Done — useful to confirm nothing
slipped between tabs. Default sort is by Status (workflow order: New → Triaged →
In progress → Blocked → Done → Parking → Won't fix). Click any row to open the item.
Phase 1 prototype.
Static data, mocked AI analysis, mocked clinician comments. No persistence — clicking
Save just shows a confirmation. Phase 2 wires Firestore
/issues and
/issues/{id}/comments, plus a Cloud Function that calls Claude/Haiku for
real AI analysis on issue create/edit. Phase 3 opens Clinical Chats to the partner role
so Rutgers clinicians can comment directly on issues that affect their practice.