2026-10-09 - 2026-11-09

Overview

8 Active Pull Requests
37 Active Issues
Excluding merges, 1 author has pushed 7 commits to main and 24 commits to all branches. On main, 37 files have changed and there have been 1818 additions and 1138 deletions.

1 Release published by 1 user

Published v1.1.0 2026-11-09 00:34:51 -05:00

8 Pull requests merged by 1 user

26 Issues closed from 1 user

Closed #302 /upload has no idempotency: a retried upload re-runs Whisper/Gemini/correlation on the same call_id 2026-11-08 21:51:18 -05:00

Closed #301 /upload: a late (spooled) upload is correlated against now, not the call's started_at 2026-11-08 21:51:16 -05:00

Closed #300 Alerts: confidence gate before any push alert (Discord/webhook) — avoid the CrimeRadar false shots-fired failure 2026-11-08 21:51:15 -05:00

Closed #104 Reprocess with stale call-doc severity: not-thin call re-extracted to zero scenes attaches with no fit check 2026-11-08 21:51:14 -05:00

Closed #162 transcript_correction.py accepted the model's rewrite verbatim — a real ten-code got swapped for a different real ten-code 2026-11-08 21:51:12 -05:00

Closed #160 place_verifier.verify() has the same no-anchor gap #159 fixed in intelligence.py's geocoder — never ported over 2026-11-08 21:51:10 -05:00

Closed #297 Transcription: move to an OpenAI-compatible cheaper provider (Groq whisper-large-v3-turbo) with a domain vocabulary prompt 2026-11-08 21:51:09 -05:00

Closed #298 Extraction: one cheap call per transcript, skip chatter before any LLM, config-driven model, cache geocodes 2026-11-08 21:51:06 -05:00

Closed #88 resolved and reassignment are prompt-only invariants with no validator, so a string "false" resolves an incident 2026-11-08 21:51:04 -05:00

Closed #83 transcript_correction_enabled is not a pure cost lever: turning it off disables the noise gate and pushes over-merging 2026-11-08 21:51:03 -05:00

Closed #138 LLM-orphan gate's "type" veto fires on visibly content-free chatter (18/22 this window) 2026-10-27 20:27:33 -04:00

Closed #91 LLM correlation tier sees the whole-call transcript for every scene, so it cannot separate scenes and pays N times for one prompt 2026-10-27 20:27:32 -04:00

Closed #134 incident_correlator fast/thin path has the same dispatch/tactical name-guess as the #115 escape hatch, but it changes link decisions 2026-10-27 20:27:31 -04:00

Closed #151 Board 2026-09-13 — Autonomous email SOP — FINAL MINUTES 2026-10-27 19:27:13 -04:00

Closed #146 Board 2026-09-13 — Beachhead checkpoint miss and dev-progress review — FINAL MINUTES 2026-10-27 19:27:12 -04:00

Closed #99 Board process defect: the 2026-08-31 sitting filed four drafts and no final minutes for 24h 2026-10-27 19:27:12 -04:00

Closed #193 Unattended runner is dead since the Linux move — Task Scheduler/run-session.ps1/drb-tg have no Linux equivalent 2026-10-27 19:27:11 -04:00

Closed #49 Board skill: CTO draft #40 was filed with the CISO draft body byte-for-byte, losing the CTO record 2026-10-27 19:27:11 -04:00

Closed #57 P0: standing correlation-quality evaluation procedure (scheduled AI window) 2026-10-27 19:27:10 -04:00

Closed #56 Telegram control plane: owner-to-agent commands — built as drb-telegram-ops (WSL), pending live verification 2026-10-27 19:27:10 -04:00

Closed #196 Prod AI dark since 09-29/09-30: Gemini and OpenAI credits exhausted (billing, not a bug) 2026-10-27 19:27:09 -04:00

Closed #66 TRACKER: 12 discovery conversations by 2026-09-22 (GOALS.md kill criterion) - checkpoint 2026-09-05 2026-10-27 19:27:09 -04:00

Closed #283 Systems page offers create/edit/delete and AI panels to operators; every backend route is admin-only, and Delete fails silently 2026-10-25 22:17:26 -04:00

Closed #222 Node page shows admin-only buttons to non-admins (Restart OP25, override Ack/Force Revert): dead CTAs 2026-10-25 22:17:26 -04:00

Closed #286 Raw 'C2 API error 403: {"detail":...}' shown to users across most screens 2026-10-25 21:39:16 -04:00

Closed #136 Map Incident Tiles are Unreadable 2026-10-25 21:03:23 -04:00

18 Issues created by 1 user

Opened #283 Systems page offers create/edit/delete and AI panels to operators; every backend route is admin-only, and Delete fails silently 2026-10-25 21:31:55 -04:00

Opened #284 Admin actions swallow failures (console.error only): Resolve, Summarize, alert rules, Ack, Restart OP25; empty states lie on load failure 2026-10-25 21:31:55 -04:00

Opened #285 Marketing CTAs contradict each other: 'Get started' -> /login, waitlist says signup closed, /signup is live 2026-10-25 21:31:56 -04:00

Opened #286 Raw 'C2 API error 403: {"detail":...}' shown to users across most screens 2026-10-25 21:32:13 -04:00

Opened #287 Settings > API Keys is an in-memory mock shown to customers; empty state claims it authenticates integrations 2026-10-25 21:32:13 -04:00

Opened #288 Node page: missing/denied node spins 'Loading node…' forever; 'No calls recorded' is wrong for quiet nodes 2026-10-25 21:32:14 -04:00

Opened #289 Stale or false copy: bot /link points to a nonexistent Settings tab; FAQ/features claim operator management and tuning persistence 2026-10-25 21:32:14 -04:00

Opened #292 Live map: stale-state dim greys out popups and legend; node pins nearly invisible on the dark basemap 2026-10-25 23:12:29 -04:00

Opened #293 Expired session renders pages signed-out instead of redirecting to login 2026-10-25 23:12:29 -04:00

Opened #296 Cut AI pipeline cost: evaluate cheaper / alternative models for STT, scene extraction, correlation 2026-10-27 19:27:14 -04:00

Opened #297 Transcription: move to an OpenAI-compatible cheaper provider (Groq whisper-large-v3-turbo) with a domain vocabulary prompt 2026-10-27 20:03:49 -04:00

Opened #298 Extraction: one cheap call per transcript, skip chatter before any LLM, config-driven model, cache geocodes 2026-10-27 20:03:49 -04:00

Opened #299 Correlation: rules first, one cheap LLM call only when ambiguous; remove the cheap/smart consensus tiebreak 2026-10-27 20:03:49 -04:00

Opened #300 Alerts: confidence gate before any push alert (Discord/webhook) — avoid the CrimeRadar false shots-fired failure 2026-10-27 20:03:50 -04:00

Opened #301 /upload: a late (spooled) upload is correlated against now, not the call's started_at 2026-10-27 20:14:56 -04:00

Opened #302 /upload has no idempotency: a retried upload re-runs Whisper/Gemini/correlation on the same call_id 2026-10-27 20:14:57 -04:00

Opened #303 call_start overwrites the call doc (merge=False) and can erase audio_gcs_uri when it lands after /upload 2026-10-27 20:38:15 -04:00

Opened #304 Correlation: make the P25 radio ID (srcaddr) a first-class link, and measure how often calls carry it 2026-11-08 01:07:48 -05:00

22 Unresolved Conversations

Open #190 Severity never de-escalates on stand-down: 'Flames at the house' stays Major/active after units report a fireplace 2026-11-08 21:40:57 -05:00

Open #189 Long-lived Major incident becomes a sink: 'Rollover Entrapment at TSP' absorbed 40 calls / 2h34m / 37 km path on Ch1 2026-11-08 21:40:57 -05:00

Open #27 Sweep retry budget can expire before the incident the orphan belongs to exists 2026-11-08 21:40:57 -05:00

Open #115 Consensus tiebreaker manufactures incidents — resolves ~all disagreements as "new", overriding the LLM's "link" 2026-11-08 21:40:57 -05:00

Open #103 content_divergence veto goes dark when the scene embedding is legitimately None (post-#80/#95) 2026-11-08 21:40:57 -05:00

Open #86 Correlation over-merge veto uses a bare 0.82 literal, looser than the 0.93 match bar it sits beside 2026-11-08 21:40:57 -05:00

Open #155 Incidents never actually clear — units_cleared almost never populates, 90-min idle sweep does the resolving instead 2026-11-08 01:07:48 -05:00

Open #35 LLM correlation tier decides only 3.5% of links, and the dominant fast/thin path records no fit signal 2026-10-27 20:03:50 -04:00

Open #242 Map: storm/utility hazards land as 'other' (?) — hazard subtype for wires down, manhole fire, flooded road, tree down, road closed 2026-10-27 20:03:49 -04:00

Open #163 STT quality initiative: raw transcription accuracy is the ceiling every downstream fix (correction, extraction, correlation) is hostage to — research + candidate angles 2026-10-27 20:03:49 -04:00

Open #48 Gate B8: transcript entity-name accuracy has never been measured, yet names are being published 2026-10-27 20:03:49 -04:00

Open #36 Move transcript correction into transcription, with per-talkgroup vocabulary and area context 2026-10-27 20:03:49 -04:00

Open #7 Bulk reprocess calls stored without a transcript 2026-10-27 20:03:49 -04:00

Open #45 Gate B5: no per-call/system/org AI cost accounting exists, so no BUSINESS_MODEL.md section 4 figure has ever been checked against a bill 2026-10-27 19:27:14 -04:00

Open #170 Admin Replay: re-run the pipeline over past calls in a sandbox, so correlation changes are measured without live AI windows 2026-10-27 19:27:14 -04:00

Open #207 No pre-merge frontend typecheck: no Node on the authoring box, no lockfile, CI only on main 2026-10-26 02:24:13 -04:00

Open #109 Frontend punch-list (drb.cusano.net live review, 2026-09-07) 2026-10-25 21:32:22 -04:00

Open #9 Member and node write endpoints are platform-admin-only, blocking org owners 2026-10-25 21:32:22 -04:00

Open #111 SDR tuning (PPM / preset) is coupled to system assignment and silently wiped on reassign 2026-10-25 21:32:14 -04:00

Open #44 Gate B4: lib/billing.ts publicly promises 7/90/365-day retention with no enforcement anywhere 2026-10-25 21:32:14 -04:00

Open #3 Billing UI is mock data and shows every org a fabricated trial 2026-10-25 21:32:13 -04:00

Open #47 Owner-only pre-revenue gates: ToS/Privacy/AUP, business entity + E&O, delay policy, agency takedown contact 2026-10-25 21:31:56 -04:00