Shared DB eprpywqwzsvrjeaforei has ZERO backup coverage: free plan, backups list empty, PITR off, no dump machinery on either tenant

TERRA-7 terra Blocked urgent

enamel_031's NOT NULL no-default on shared supportTickets broke requestTerraAccess for 9 minutes: zero eligible callers, nobody hit it

TERRA-40 terra Done urgent

PROD debe ser copia del sqlite de RUTAP (v_rutap_latest)

MOVILBA-13 movilba Backlog urgent

└─ CARGAR: 4 dominios EN SERVICIO ausentes de prod (PMH988 NO entra, va en MOVILBA-24)

MOVILBA-18 movilba Done urgent

└─ REACTIVAR 11: 1 dominio soft-deleted en prod y vigente en RUTAP (PMH988)

MOVILBA-24 movilba Done high

└─ BUG: normalizar_dominio borra minusculas en vez de normalizarlas

MOVILBA-27 movilba Done high

└─ CORREGIR 7: 2 celdas ################ visibles en el export de Eugenia

MOVILBA-20 movilba Done high

└─ HARD DELETE: 20 entidades de prueba + linea 945 id 176

MOVILBA-32 movilba Done high

└─ PROD desactualizado: Atlantida asignada a linea 945 cerrada, RUTAP la tiene en linea 0

MOVILBA-29 movilba Canceled high

└─ DEFECTO: mismo texto de chasis mapeado a dos chasis_id distintos

MOVILBA-28 movilba Canceled high

└─ ASIGNACIONES: diff bidireccional prod <-> sqlite

MOVILBA-17 movilba Canceled high

└─ VEHICULOS: diff bidireccional prod <-> sqlite

MOVILBA-16 movilba Canceled high

└─ LINEAS: diff bidireccional prod <-> sqlite

MOVILBA-15 movilba Canceled high

└─ ENTIDADES: diff bidireccional prod <-> sqlite

MOVILBA-14 movilba Done high

└─ CORREGIR 12: clase F, conflictos reales campo a campo

MOVILBA-25 movilba Backlog high

└─ CORREGIR 9: diff a nivel de campo de los 2260 dominios que matchean

MOVILBA-22 movilba Done high

└─ CORREGIR 10: importador debe ingerir altas recientes y filas cerradas

MOVILBA-23 movilba Backlog high

└─ CARGAR 6: 131 dominios cerrados en RUTAP nunca importados

MOVILBA-19 movilba Done

└─ CORREGIR 13: 8 dominios con mas de una fila en vehiculos

MOVILBA-26 movilba Done

└─ CORREGIR 8: linea 945 duplicada en prod

MOVILBA-21 movilba Done

└─ Reclasificar los 12 pares de operador real que MOVILBA-17 dio por historia legitima sin medirlo

MOVILBA-31 movilba Done

└─ Relleno '0' textual en id_adas / ip_mdvr / ip_canbus: 2012 celdas, cero NULLs en prod

MOVILBA-30 movilba Backlog

crearPracticaFailed INSERT column/value mismatch

VENUS-244 venus Done urgent

Unattributed full-DB reader on venus Supabase project (~3/day, postgres role, incl. vault.secrets) — attribute or treat as compromise

VENUS-350 venus Blocked urgent

Audit: session-trust paths vs isGenuineGoogleSession (auth-bypass check)

VENUS-304 venus Done urgent

SEC: portal server actions ungated - unauthenticated read of client queue

ENAMEL-18 enamel Done urgent

URGENT: fix quoted search_path in mig 099/102 -- 33 functions broken, prod outage since 13:03 UTC

PLUTO-673 pluto Done urgent

applog-listen: a log-only row PERMANENTLY PINS the cursor (classify inside void enqueue) — venus dead 15 days, latent on all 4 rails

VENUS-324 venus Done urgent

applog-listen sends to an ARRAY-of-one and reaches the UNGUARDED hub resolver: scalar to + human-seat recipient validation + regression (EVO-86(e))

VENUS-323 venus Done urgent

chat-duo (cdw) Pending pane clobbers focused input on refresh-poll

MSG-259 llmmsg-srv Canceled urgent

chat-duo Elazar Pending pane refreshes mid-type → loses user input. Bug: pane re-render clobbers focused input field. Owner: llmmsg-srv-cc-w (chat-duo).

MARS-290 mars Done urgent

Práctica date off-by-one: stored/displayed -1 day from entered

MARS-286 mars Done urgent

CRITICAL: cannot log in (client report via pluto)

AYUD-76 ayudarg Done urgent

P0: validate ids, status, and priority in wi and wi-web

BSQ-2 basquet Done urgent

docentes/[id] still RSC-crashing post-v1.82.1 (digest 3811385828, eventId 5401bdaf)

MARS-263 mars Done urgent

SMTP fully broken: SMTP_HOST env var has literal \n suffix -> EDNS getaddrinfo fail, ALL outbound email fails (notify/review/migration). Fix: defensive .trim() on SMTP_* env reads in mailer config + correct Vercel env vars.

MARS-319 mars Done urgent

test urgent with agent

OPS-39 agent-ops Canceled urgent

BUG: número de documento input invisible on /practicas/nueva (CSS regression from v1.54 form-field width fix)

MARS-223 mars Done urgent

P0: wrap wi and wi-web writes in transactions

BSQ-1 basquet Done urgent

Live 500 on /informes — fn_getInformesBundle returns porAnestesia (and likely sibling array keys) undefined for sparse/zero-row scope; client .map crashes. PLUTO-97 regression. Fix fn to COALESCE every array key to [] + client null-guard. PTD must GET /informes for the failing scope, not log-tail.

PLUTO-112 pluto Done urgent

MSG-76: /setpm error feedback — DM sender when command fails (wrong arg count, unknown agent, not authorized)

MSG-76 llmmsg-srv Done urgent

Backfill KEY-N refs (projectKey+projectNumber) for all non-deleted items — BWI-21 prerequisite

BWI-22 basquetwi Done urgent

Access regression: students who previously logged in are now blocked (agujero negro / blocked-user landing) - PROD, real users

PLUTO-69 pluto Done urgent

SECURITY: rotate pm_triage_reader DB password (committed literal in git history, PLUTO-191 GRANT migration)

PLUTO-194 pluto Done urgent

URGENTE 4h: comisión + JTP REQUIRED en /solicitar-acceso (rama estudiante) — submit bloqueado si faltan

PLUTO-349 pluto Done urgent

PROD-DOWN: práctica create fails — studentAssignmentsArchive trigger fn still INSERTs dropped isDemo column (MARS-78 class, slipped table)

MARS-88 mars Done urgent

EPIC: ONEDB target flip — telemetry consolidation onto venus-pg (supersede whey-pg)

VENUSINF-1 venusinfra Backlog urgent

└─ Confirm + record: venus-pg is the consolidation target; supersede whey-pg in ONEDB-12/10 + one-db/DESIGN.md

VENUSINF-6 venusinfra Backlog

└─ Repoint venus-kpi-collect + whey/lezama remote writers to venus-pg

VENUSINF-9 venusinfra Backlog

└─ Provision postgres17 on venus (data dir, listen_addresses, telemetry db + telemetry_rw role)

VENUSINF-7 venusinfra Backlog

└─ pg_hba + network: allow whey + lezama as remote writers to venus-pg

VENUSINF-8 venusinfra Backlog

└─ Cross-update ONEDB WIs (ONEDB-12 ingestion, ONEDB-10 pg-write flip) for new target host

VENUSINF-10 venusinfra Backlog

URGENT: app logger dead since 2026-05-28 — restore logging (logs table) so we have troubleshooting data

AYUD-1 ayudarg Done urgent

Grading authz audit: práctica-grade permission scope vs Puia rule (comisión-wide adjunto/jtp/ayudante + titular-all)

PLUTO-141 pluto Done urgent

fn_getInformesBundle: include 0-practica JTPs (Opt A) across jtpStats + comparativaJtp + jtpBenchmark

PLUTO-144 pluto Done urgent

URGENTE 2h: post-login redirect a form de auto-asignación para alumnos sin comisión/JTP — auto-aprueba selección

PLUTO-346 pluto Done urgent

URGENTE 2h: email a alumnos huérfanos con link a /completar-asignacion

PLUTO-347 pluto Canceled urgent

URGENTE 4h: /solicitar-acceso para estudiantes incluye pulldown comisión + JTP + datos esenciales

PLUTO-348 pluto Done urgent

Nav visibility gap: /practicas/nueva + /mi-cursada unreachable via sidebar for non-alumno roles

TERRA-17 terra Blocked high

Retire/reactivate writes archive rows with no actor or reason: RETIRE_SQL runs on a bare pool.query so setArchiveContext cannot reach it, and terra_29 will break the feature outright

TERRA-78 terra Todo high

Modulos offer AI-seeded cirugias instead of the catedra list: client-reported, correct mapping supplied, Drive source not yet in hand

TERRA-75 terra Blocked high

Implantologia and reconstruccion alveolar need modulo-specific fields (injerto, zona dadora, barrera, marca de implantes): never built, shape undecided

TERRA-76 terra Backlog high

route-dedup.ts: two open unowned defects (windowing collapse, unrecoverable summary)

VENUS-375 venus Done high

A probe row in appEvents is indistinguishable from a real user failure and the record is permanent

TERRA-43 terra Blocked high

Guard suite unexercised on a non-empty population: the writer exists and is owned, but no natural terraPracticas write has been observed

TERRA-46 terra Todo high

The citation gate checks tag SHAPE, never tag REALITY: a well-formed invented token passes exactly as a genuine one does, and reads to a human as a real citation

TERRA-69 terra Done high

Reason codes assert the data is wrong and terra has no correct-or-retire path for a cirugia: TERRA-55 risks showing staff a fault they cannot act on

TERRA-56 terra Done high

Vocabulary-table archive-on-UPDATE: terraCirugias has an INTEGER PK against a uuid NOT NULL sink, so the generic archive function cannot be reused and the sink assumption itself is in question

TERRA-66 terra Todo high

portalPen: staff triage surface (read, answer, assign, resolve tickets)

ENAMEL-20 enamel Done high

soporte-action agent CLI was DEAD: three route/client drifts on the headless support path

VENUS-373 venus Done high

└─ Smoke test: exercise scripts/soporte-action.ts against the live support-action contract

VENUS-374 venus Done low

Raise with applog-listen.ts owner: refuse startup when co-tenanted DB has APPLOG_APPKEY unset

TERRA-25 terra Done high

Client cannot confirm a cirugia: isConfirmedByClient has 3 readers and no writer, and it blocks every guard on the empty population

TERRA-52 terra Done high

Support-ticket surface fails the agent-operable directive: mutation and thread capabilities are MISSING, not merely auth-unreachable

TERRA-36 terra Blocked high

Shared support tables have zero DB-side access control: no RLS, owning role, cross-tenant with terra

ENAMEL-28 enamel Backlog high

Agent-operable ticket API: REST handlers over /soporte, single auth definition

ENAMEL-27 enamel Blocked high

Session dies mid-session on read-only navigation ~1h after login

TERRA-22 terra Blocked high

Raise coder/coder02/db/db02 auto-compact window to 400k (CLAUDE_CODE_AUTO_COMPACT_WINDOW in each agent's own .claude/settings.json)

MARS-644 mars Done high

Actuator tiers must be per-agent-window: scale absolute gates (HARD_NUDGE_K/HARD_COMPACT_K, quarantine) by each agent's own CLAUDE_CODE_AUTO_COMPACT_WINDOW instead of assuming one fleet window

OPS-128 agent-ops Done high

Support-ticket API parity - create/triage/status ops + list read, no UI-only actions

PLUTO-700 pluto Backlog high

Terra hard-delete guard coverage: archive ledger, 4 clinical detail tables, counter

TERRA-10 terra Done high

terraPacienteImplantologia: patient-clinical table with deletedAt and apparently no archive trigger on any mutation

TERRA-31 terra Done high

Detail-table UPDATEs bypass both archive triggers: clinical content overwritable with zero trace

TERRA-28 terra Done high

Venus-consolidation: unified service on host venus for Vercel/Supabase background work

EVO-86 evolutiva Backlog high

Mars OAuth redirect_to bypasses resolveSiteOrigin() + 3 uncanonicalized prod hosts (TERRA-20-derived)

MARS-642 mars Backlog high

Unprotected .vercel.app project alias serves the full production build, and is the PKCE-mismatch mechanism

TERRA-20 terra Done high

AUDIT dropbox-orphan-sweep.sh (registry row 116): what it deletes, how it fails

OPS-127 agent-ops Done high

Interno desactualizado en 2 vehiculos: AB674IP y AF106VS no coinciden con el SSOT de RUTAP

MOVILBA-43 movilba Backlog high

Refactor: historial de vehiculo pasa a log append-only con workflow (reemplaza el parche de forzar Otro)

MOVILBA-42 movilba Backlog high

└─ vehiculo_historial: no hay columna que diga cual es la fila vigente - borrado_en y creado_en estan sobrecargados

MOVILBA-12 movilba Backlog

MOVILBA-34 parte 2: llevar expediente_baja a PROD (DDL + vista) y backfillear 2440 filas cerradas desde RUTAP

MOVILBA-36 movilba Done high

Hub must not return a success that asserts more than it verified (ok:true vs delivered)

MSG-285 llmmsg-srv Done high

└─ Wedged/booting session is electable as PM and can be handed the queue flush (register fakes liveness)

MSG-289 llmmsg-srv Backlog urgent

Hub response contract: ok:true must not assert more than was verified (pm_pending_queue unbounded + resolved_pm null on reroute)

MSG-282 llmmsg-srv Canceled high

User + group administration: no UI exists, every grant is a hand-written DB row

ENAMEL-22 enamel Blocked high

Live schema and db/migrations have drifted: full both-directions inventory needed

ENAMEL-26 enamel Backlog high

Access-request form: /solicitar-acceso is a dead end, accessIssueReports never written

ENAMEL-21 enamel Done high

Decide hosting for the new websites-tools git repo

WEBSITES-2 websites-tools Done high

Most sites have no GA4 property at all (16 of 49 covered)

WEBSITES-8 websites-tools Deferred high

Grant service account GA4 Administrator on all three accounts

WEBSITES-10 websites-tools Done high

Agents running outside tmux are silently uncoverable by the actuator: no statusline to read, no pane to keystroke, no alert (pm-llmmsgsrv-cc at 76.4% uncovered right now)

OPS-125 agent-ops Backlog high

cc-context-monitor last_turn() v2.14 residual stale-read: agent force-compacted 'for no reason' every other session. Verified 2026-06-09 nw-whey-cc reported 184k while live /context=63.6k (3x phantom). last_turn (~line 312) reads transcript's last usage block = summarizer's pre-compact cache_read until next real turn; compact_boundary reverse-walk skip has a hole (boundary outside TAIL_BYTES, or pre-compact reading sampled in the gap before boundary written). Fix: gate actuator on LIVE window occupancy, not stale transcript estimate. NOT a cap problem - do NOT bump nw ROLE_CAP (Elazar explicit). Recurring user-facing churn.

OPS-11 agent-ops Done high

CronCreate /compact unreliable in never-idle loop: past-dated one-shots persist as zombies (363-day timer), monitor DM injection prevents REPL idle — path-B prose /compact is inert

MSG-89 llmmsg-srv Done high

tick.py execute() is sequential: one slow agent blinds the whole roster and delays post-nudge escalation (2.5min measured)

OPS-106 agent-ops Done high

to:[] multi-DM array path never inherits re= parent origin_aro

MSG-279 llmmsg-srv Done high

Post-compact state is structurally unmeasurable: 0k filter hides landed compacts, so compact_unverified fires on successes

OPS-112 agent-ops Done high

No regression guard on the idle classifier: a ghost-vs-typed inversion caused a silent fleet-wide coverage outage

OPS-123 agent-ops Done high

Stuck-busy agents leave coverage silently: unsubmitted prompt text pins idle=false, 14h measured uncovered, no alert

OPS-116 agent-ops Done high

abr:ss attribution false-negative: fast responders fail the text_in_pane visibility test

OPS-121 agent-ops Done high

nudge_pending_since consumed on decide, not on confirmed exec: an aborted escalation can shut all four triggers

OPS-108 agent-ops Done high

Persist-marker latch: PRE-COMPACT PERSIST re-asks an agent that already replied DONE-NO-DELTA (4 confirmations, needs agent_state migration + cross-rail state)

OPS-115 agent-ops Backlog high

Copia staging->dev de vehiculos: dev arranca de cero con datos reales

MOVILBA-1 movilba Done high

└─ Anuncio en aro:mba-l + modo mantenimiento en dev

MOVILBA-4 movilba Done

└─ Verificacion post-carga

MOVILBA-7 movilba Done

└─ Secuencias (trampa 23505 de archive)

MOVILBA-6 movilba Done

└─ Carga transaccional

MOVILBA-5 movilba Done

└─ Ensayo de rollback

MOVILBA-8 movilba Done

└─ Backup dev-before + checksums

MOVILBA-3 movilba Done

└─ Preflight: probar que los esquemas son identicos

MOVILBA-2 movilba Done

└─ Purga de logs fresh-start (app_logs + audit_log)

MOVILBA-9 movilba Backlog

Operator-alert attempt ceiling (3.5s) sits below the SMTP transport's own connect timeout (10s) — races away connections that would have succeeded

PLUTO-688 pluto Backlog high

P0: fn_insertSupportTicket ambiguous id column broke all support-ticket creation

PLUTO-677 pluto Done high

Email drain: kill ~288/day cron no-ops (Tier 0 EXISTS-guard) + DB-native trigger-wake (Tier 1)

EVO-70 evolutiva Done high

└─ EVO-70 Tier 1: emailOutbox AFTER INSERT FOR EACH STATEMENT trigger → pg_net wake marker to drain route

MARS-520 mars Done

└─ EVO-70 Tier 1: trigger-wake for emailOutbox (pg_net)

PLUTO-613 pluto Done

└─ EVO-70 Tier 0: guard drain-email-outbox cron with EXISTS check

PLUTO-612 pluto Done

└─ EVO-70 Tier 0: conditionalize drain-email-outbox pg_cron — EXISTS-guard so empty ticks fire zero HTTP/Vercel

MARS-519 mars Done

└─ MARS-519 follow-up: add drift-breadcrumb comment in email-outbox.ts tying RECLAIM_AFTER + claim eligibility to the pg_cron T0 guard fn

MARS-521 mars Backlog low

└─ EVO-70 Tier 1 follow-up: retire app after() calls (gated on passive prod evidence)

PLUTO-615 pluto Blocked low

isUuid rejects legacy deterministic seed UUIDs (non-RFC version/variant nibbles), false-404s comisión/other routes

PLUTO-578 pluto Done high

fn_notify_applog_err collapses all 404s into one signature bucket

VENUS-243 venus Done high

Rectificaciones fix: (A) DB sweep 10 stuck-rectificar comA practas to noRevisado, (B) exempt rectificar from review dedup so re-rectificar works, (C) Reenviada-tras-rectificar badge on noRevisado rows, (D) teacher notification email on student resubmit

MARS-316 mars Done high

Solicitudes de acceso: role (student vs teacher) not visually clear — Estudiante tag present but teacher-like fields (Cátedra postgrado, JTP/Ayudante) rendered alongside, confusing reviewer

MARS-315 mars Done high

Ayuda y Soporte: in-app support-ticket flow — global FAB + Dialog + supportTickets table + internal-email notify Elazar on creation + Vercel cron hourly digest DM to PM + watchdog open-count poll (per Elazar -36580 + pm-mars-cc design -36593)

MARS-312 mars Done high

Minimize non-actionable PM DMs at the ALERTER (not just app-side rows). Each alert DM = a PM turn = tokens (Elazar 2026-06-29). Scanner-probe silencing cut the warn ROWS fleet-wide (PLUTO-450/453/454 + mars/venus), but the applog/errscan digest still DMs each PM per tick incl non-actionable warn/info lines. RULE: alerter DMs a PM ONLY on level=error/fatal OR a genuinely NEW/unseen signature; NEVER re-DM known silent-tier (info/security) scanner shapes or recurring-benign signatures. Two legs: (1) shared errscan digest (owner bin-whey-cc, ~L244/L309 PM-digest render + the source view predicate) suppress non-error/non-novel; (2) per-project applog-pull digest (each PM's project) same predicate. Fleet-wide; all 3 PMs align in aro:evolutiva-management. Distinct from EVO-46 (count mislabel) though same render path.

EVO-47 evolutiva Backlog high

SOL-VENUS-N human reference for access requests

VENUS-360 venus Done high

Enumerate legitimate DB principals for Mars (fleet DB-hardening step 2)

MARS-620 mars Done high

venus-db-backup.sh: trash||rm precedence hard-deletes clinical dumps when trash exists and FAILS

VENUS-359 venus Blocked high

venus-db-backup.sh: prune glob spans plaintext+encrypted lanes, sharing one KEEP budget

VENUS-358 venus Blocked high

Enable log_connections (or equivalent) on Mars Supabase DB

MARS-619 mars Done high

Consolidate Evolutiva backups into one performed/monitored/inventoried lane

VENUS-357 venus Backlog high

pm-triage-listen: rowless NOTIFY kinds (weeklyReport) silently dropped

VENUS-353 venus Done high

Fix /administrar/usuarios throwing past 500 active users - add server-side pagination

MARS-615 mars Done high

SOP-MARS-39: período endDate silently blocked ALL práctica filing for 2026-1C-Grado cohort since 2026-08-01 (4-day undetected outage)

MARS-607 mars Done high

Materia-dimension migration 099/100: setActivePeriodo cross-materia blackout + createPeriodo missing materiaOptionId

MARS-606 mars Done high

SMTP host/port/secure validating resolver — replace the || default on the email-send path

MARS-601 mars Done high

Terra outbound mail: access-request notification (NOT migration terra_13 — that is TERRA-10's deliverable)

TERRA-13 terra Done high

hub_send.py marker path was __file__-relative: correct only because the unit names the symlink, and it reported landed registrations as failures

OPS-114 agent-ops Done high

Idle prefetch used the 50% growth floor and compacted agents with 35pt of headroom (coder-pluto-cc at 51.2%)

OPS-113 agent-ops Done high

Foto de perfil obligatoria para alumnos + foto del alumno en la UI de corrección

MARS-600 mars Done high

Idle-path compact runaway: NULL growth baseline made every above-floor agent perpetually eligible (pm-terra-cc 8x in a morning)

OPS-111 agent-ops Done high

deletedAt on the four terra clinical detail tables (ruled 2026-07-31, not delivered by terra_13)

TERRA-15 terra Done high

Portal: accept 'venus' as a portal appKey scope tag + grant Puia client under venus

ENAMEL-17 enamel Done high

Repoint 5 remaining APP_URL fallback sites to resolveSiteOrigin()

MARS-599 mars Done high

Greenfield transactional email stack: dispatcher chokepoint, admission+egress guards, durable outbox

ENAMEL-16 enamel Backlog high

cc-context-monitor over-cap caps are stale — false alarms fleet-wide. THRESH_OPUS=160k/HAIKU=100k/SONNET=135k are ~80% of the OLD ~200k-class windows. Fleet now runs claude-opus-4-7[1m] = 1M-token window; a 268k session is 27% full, not over cap. Transcript model string lacks the [1m] marker so monitor can't auto-detect window size. Fix: bin-whey-cc updates THRESH_* in cc-context-monitor.sh to real values per Elazar's decision (true-window fraction vs. relabelled cost-budget). All of 2026-05-21's over-cap wind-downs were triggered by this false cap.

OPS-53 agent-ops Blocked high

Venus DB backup lane: executor unknown, last local dump 2026-06-16 — identify owner or declare backup-gap incident; fix db/README.md §Backup

VENUS-348 venus Done high

6 userRoles call sites bypass soft-delete check -- role grant/revive + last-admin-guard defeat on deleted users

PLUTO-640 pluto Done high

Data API OFF is an unowned config that terra's pg_temp LATENT rating silently depends on

TERRA-11 terra Done high

Split shared app.bypass_immutable GUC per-guard (enamel half)

ENAMEL-13 enamel Done high

userRoles last-admin trigger missing ENABLE ALWAYS (replica-mode bypass, same class as mig 101)

PLUTO-671 pluto Done high

getStudentsWithoutAssignment missing ur.deletedAt IS NULL -- 6 of 7 rows on the live student-assignment picker are reclassified docentes/admins, not students

PLUTO-667 pluto Done high

Land PLUTO-641 orphaned app-side refactor (admin fn_* calls)

PLUTO-665 pluto Done high

Unpinned authz-scope functions need explicit search_path design (pg_temp fleet finding)

PLUTO-666 pluto Backlog high

Whole-DB pg_dump insurance for eprpywqwzsvrjeaforei (enamel+terra shared project) does not exist

ENAMEL-12 enamel Blocked high

Fleet CR + client-questions surface (replace Google Sheets/Docs for client change-requests and our questions-to-client)

EVO-83 evolutiva In progress high

└─ Terra leg of EVO-83: CR + client-questions surface

TERRA-1 terra Done high

└─ Adopt fleet CR + client-questions surface in venus (EVO-83 leg)

VENUS-317 venus Done

applog rail: 404/probe path CLASSIFIER + 4-outcome taxonomy — must land BEFORE any suppression predicate on any rail (both per-app predicates withdrawn)

VENUS-327 venus Backlog high

applog-listen category filters assume VENUS literals — both are INERT on pluto (security=0 rows, access=0 rows); four rails, one file, four vocabularies

VENUS-326 venus Backlog high

getFrictionDigest: column reference "createdAt" is ambiguous — real SQL error, invisible 14 days because category='security' is bot-filtered to log-only

VENUS-325 venus Backlog high

EVO-86: pluto applog rail — real error pages self-addressed to scrp-applog-pluto, never reached pm-pluto-cc

PLUTO-652 pluto Done high

applog alert delivery: pages routed to the emitter script (scrp-applog-*) instead of a human inbox, all rails

VENUS-322 venus Backlog high

Push staged appEvents-standard + retired-identifiers commons edits once objection window closes (hub 07:58 / 10:58 UTC 2026-07-30)

EVO-85 evolutiva Backlog high

Alert-rail per-app routing for terra errors

TERRA-2 terra Canceled high

auth/callback 42P08: raw bind-placeholder in accessRequestMatchExpr blocks real logins

PLUTO-647 pluto Done high

support-actions.ts stamps persona id instead of realUserId in actor/archive/log fields

PLUTO-645 pluto Done high

Retroactive audit-PASS verification: 21 Class-A commits with uncited self-noted verdicts

PLUTO-646 pluto Backlog high

support-actions.ts stamps persona id instead of realUserId in actor/archive/log fields

PLUTO-644 pluto Canceled high

Land or discard dangling postgres-boundary WIP (migrations 089-094)

PLUTO-641 pluto Done high

Postgres-boundary WIP: tree calls 4 DB functions that do not exist live; migrations 089-094 unapplied

PLUTO-642 pluto Canceled high

VENUS-305 BLOCK fix-forward: incomplete identityBound writer coverage

VENUS-311 venus Done high

getAcademicDropdownData anonymous-callable, same roster-PII leak class as PLUTO-635

PLUTO-638 pluto Done high

Audit: roleless-but-sessionful paths vs genuine-Google gate

VENUS-306 venus Done high

Supabase open email/password signup — disable external_email_enabled

VENUS-302 venus Done high

/solicitar-acceso GET not genuine-Google gated (asymmetric with SUBMIT)

MARS-573 mars Done high

Disable Supabase Email provider (external_email_enabled:false) + committed authconfig.json + diff-on-push assertion

PLUTO-630 pluto Done high

Disable GoTrue email/password provider (external_email_enabled)

MARS-570 mars Done high

GoTrue-config verify-script via Supabase Management API (read-only, field-selected)

VENUS-303 venus In progress high

Remove dangling Vercel preview-alias from GoTrue uri_allow_list (evolutiva-m-pensanta-com.vercel.app)

MARS-569 mars Done high

image-proxy profile-image gate not widened for /usuarios peer directory — students get broken avatars

MARS-558 mars Done high

review:locked-by-student strands práctica ungradeable on abandoned student edit (no lock TTL/override)

MARS-547 mars Done high

Retire fn_lifecycleInactivity date-auto-disable path + clear inactivity flags (live-incident remediation)

PLUTO-627 pluto Done high

└─ Remove inactivity-lifecycle source refs (auth/callback reset + poll-security-alerts user email routing) — PLUTO-627 code tail

PLUTO-628 pluto Done high

└─ Drop 3 orphaned inactivity lifecycle columns from users

PLUTO-629 pluto Done

auth/callback rejects dangling supabaseAuthId as takeover instead of rebinding — 79 legacy users latent-locked-out

MARS-548 mars Done high

Account access-state (Activa/Desactivada), no-soft-delete-for-inactivity

EVO-81 evolutiva Backlog high

Eradicate demo-data concept residue to zero (bs-mrteu5nvtt6)

PLUTO-624 pluto Done high

Mars demo-concept eradication: sweep-to-zero + ban-guard (bs-mrteu5nvtt6)

MARS-542 mars In progress high

EVO-79 mars leg: access-request dedup PREVENTION — attach Google identity to existing pre-created account on confident match, not a 2nd user

MARS-532 mars Done high

Rename práctica date label to "Fecha de la cirugía" + show load date "Cargada el"

VENUS-296 venus Done high

Fleet-wide Gmail dot/plus normalization: block duplicate accounts at the DATA LAYER (Mars+Pluto+Venus)

EVO-45 evolutiva Done high

Supabase supautils force-grants PUBLIC EXECUTE on new public objects: fleet REVOKE-explicit standing pattern + live-exposure sweep

EVO-71 evolutiva Backlog high

└─ EVO-71 pluto leg: supautils PUBLIC-EXECUTE sweep + object-class probe + fn-migration REVOKE policy

PLUTO-614 pluto Done high

└─ EVO-71 mars: sweep live public fns for residual PUBLIC/anon/authenticated EXECUTE (supautils force-grant bypasses ALTER DEFAULT PRIVILEGES) + object-class probe

MARS-522 mars Done high

└─ EVO-71 tooling reference: author fn-REVOKE enforcement — (1) .gitpush-pre.sh static guard flagging CREATE [OR REPLACE] FUNCTION lacking same-tx REVOKE EXECUTE FROM PUBLIC/anon/authenticated + (2) post-apply catalog assertion (proacl excludes PUBLIC/anon/authenticated on new public fns)

MARS-523 mars Done

EVO-68 Mars leg: replace email-DOMAIN login gate with genuine-Google-identity gate in auth/callback

MARS-518 mars Done high

sexo: add 'No especificado' opt-out fleet-wide (keep required)

EVO-67 evolutiva Done high

└─ EVO-67: add 'No especificado' opt-out value to sexo/sexoPaciente lookup group

VENUS-280 venus Done

Restore reportedPracticeDate required + drop today-default (reverse MARS-509 over-relaxation) + add to EVO-66 flash set

MARS-515 mars Done high

Fix /perfil upload double-submit (button not disabled during await)

PLUTO-599 pluto Done high

Fix ERR_DLOPEN_FAILED sharp crash on POST /perfil

PLUTO-596 pluto Done high

Create bedel account (magali.ayelen020@gmail.com) with /estadisticas access

MARS-441 mars Done high

SOP-27 follow-up: JTP-scoping/duplication/attribution/metrics bugs across /alumnos /practicas /informes /comisiones

MARS-471 mars Done high

└─ JTP-of-whom rebuild from práctica picks (Q2 systemic reconcile) + own-students-first visibility (Q1) — Elazar-ruled

MARS-474 mars Done high

└─ Informes differing-cells nondeterminism: same comision/URL renders different aprobacion/latencia/alumno-count across viewers minutes apart

MARS-473 mars Done

Slow-query regression: json_agg lookup-label query mean 5.9s

MARS-484 mars Done high

EVO-54 phase 1: demo/seed data debt inventory

MARS-481 mars Canceled high

getStudentsForTeacher/getStudentsForTeacherPage: fact aggregates computed before roster scope

PLUTO-576 pluto Done high

EPIC: read/analysis/consumption surface

WEBUSAGE-2 webusage Backlog high

lastPractica/inactiveStudents: persisted-attribution scope-gate

PLUTO-575 pluto Backlog high

Fleet lint matchBrace bug: no-silent-failure guard undercounts broken sites (destructured params / Promise generic returns)

EVO-53 evolutiva Done high

Adopt fleet no-silent-failure lint (scripts/lint-no-silent-failure.mjs) — net-new for Pluto, not a re-vendor

PLUTO-570 pluto Done high

Hoist impersonation guards into try (silent-fail retrofit)

VENUS-251 venus Done high

Sharp native module (libvips) fails to load on Vercel linux-x64 at /perfil — ERR_DLOPEN_FAILED, breaks image optimization/serving

PLUTO-569 pluto Done high

fn_notify_applog_err 404 signature collapse — fleet-wide alert-dedup bug

MARS-439 mars Done high

Harden cron_auth_header(): REVOKE EXECUTE FROM PUBLIC + lock search_path

MARS-380 mars Done high

Approving docente_ayudante_primera access requests never inserts jtpAyudantes row

PLUTO-566 pluto Done high

bwi inter-site: Postgres + UTCP migration

BWI-23 basquetwi Done high

└─ DB: port schema SQLite-17 -> Postgres 17 (CHECK/FK kept, payloadJson -> jsonb, event-row invariant preserved)

BWI-24 basquetwi Done high

└─ DB: port live data SQLite -> Postgres + cutover (preserve ids/timestamps/events, quiesce writes, verify counts, flip source of truth)

BWI-30 basquetwi Done high

└─ Coder: thin HTTP service over Postgres enforcing bwi invariants (event rows, parentId cycle check, payload/updatedAt)

BWI-27 basquetwi Done high

└─ Ops: daemonize wi-http-server on whey (systemd unit, autostart on boot, BWI_HTTP_KEY/env mgmt, bind 127.0.0.1)

BWI-31 basquetwi Done high

└─ DB: Postgres remote access for venus/lezama (pg_hba + TLS or SSH tunnel; lean config)

BWI-25 basquetwi Canceled

└─ Coder: UTCP manual at /utcp describing the HTTP service (manual_version, tools, call templates)

BWI-28 basquetwi Done

└─ Ops: make Postgres persistent on whey (unmask postgresql wrapper + enable on boot)

BWI-26 basquetwi Done

└─ Coder: migrate wi CLI + wi-web DB driver sqlite3 -> libpq/psql after schema lands

BWI-29 basquetwi Done

Triage RSC render error on /practicas/nueva (digest, held for pause)

VENUS-258 venus Done high

Register evolutiva-errscan@venus systemd instance (currently not running)

VENUS-253 venus Done high

SOP-MARS-27: resolve support ticket — student JTP-pick secret

MARS-469 mars Done high

Extend support-api to full ticket CRUD (status, private comments, notify-list) - EVO-52

PLUTO-564 pluto Canceled high

SOP support-ticket API: full CRUD + private stakeholder comments + notify-list, per app

EVO-52 evolutiva Backlog high

└─ SOP support-api: add ticket CRUD (close/status/notify) + private stakeholder comments

VENUS-252 venus Done urgent

└─ SOP support-api: add ticket CRUD (close/status/notify) + private stakeholder comments

PLUTO-563 pluto Done high

└─ SOP support-api: add ticket CRUD (close/status/notify) + private stakeholder comments

MARS-467 mars Backlog high

Runtime/toolchain + dependency upgrade baseline (fleet-wide reference)

MARS-449 mars Done high

Fix /administrar dup-title+stale-sidebar bug; extend header/breadcrumb pattern fleet-wide across admin section

MARS-450 mars Done high

Runtime/toolchain upgrade — Next stable, React 19.2.7, Node 24.x, pnpm 10 + staged compat bumps

VENUS-245 venus Done high

pruneStale wipes aro memberships of sleeping-host agents; make membership sticky across stale-prune

MSG-270 llmmsg-srv Backlog high

Support ticket: /mi-cursada issue (registered user)

PLUTO-543 pluto Done high

Harden Postgres against trigger-bypassing hard-deletes on practicas + add deletion-integrity detection

MARS-435 mars Done high

Create vOpenSupportTickets view (fleet-wide open-ticket view standard)

PLUTO-542 pluto Done high

Recover content of 2 untraced hard-deleted prácticas (f223ef93, cc487cbb) from pre-deletion backups

MARS-436 mars Done high

Harden status-ping reroute: catch leading-*** ceremony banners to Elazar

MSG-269 llmmsg-srv Backlog high

Drop hub cold-start announce fan-out (nothing consumes it, wakes every eng-aro member on each bounce)

MSG-268 llmmsg-srv Done high

Hub: brainstorm can leave an aro frozen for hours — add auto-close/expiry so a stale session can't block new work

MSG-260 llmmsg-srv Done high

brainstorm must never hang on an offline expected-replier: auto-skip + guaranteed close

VENUS-234 venus Done high

Exempt rail/alert-sender identities from aros-prune

MSG-267 llmmsg-srv Done high

Mars alert rail blind: applog-mars pruned from aro:mars, error pages silently dropped

MARS-431 mars Done high

vFrictionEvents LIMIT 1000 truncates human admin logs dashboard - add all-time uncapped view

MARS-427 mars Done high

busy-nudge: de-gate from humanincdw + cover lane-context 1:1 asks

VENUS-225 venus Done high

humanincdw-gated suppression of worker status-pings to Elazar

VENUS-222 venus Done high

hourly LLM concision sampler for agent-to-agent messages

VENUS-224 venus Backlog high

site-config.ts localhost-fallback prod bug

VENUS-220 venus Done high

Feedback/friction telemetry — centralized instrumented capture contract (root fix)

VENUS-213 venus Done high

Friction-telemetry unification: one feedback chokepoint, canonical category=toast+mechanism, server-action transport, complete logs friction view

MARS-419 mars Done high

MARS-416 venus default-grants revoke

VENUS-212 venus Done high

Security: revoke default anon/authenticated SELECT grants across Mars public schema (83 relations incl appEvents family)

MARS-416 mars Done high

Route hub-originated alerts (silence-detector/pmidle/dormant-PM) OFF HUMAN_AGENT to a PM/ops lane

MSG-265 llmmsg-srv Backlog high

elazar-contact durable must-reach backstop: hub liveness + unread-age email fallback

MSG-264 llmmsg-srv In progress high

Solo-agent initiative: unified fleet protocol + named driving maintainer

EVO-50 evolutiva Done high

db-size-monitor.service failing exit-127 daily (orphaned script)

OPS-104 agent-ops Done high

Admin usuarios RSC render error — GROUP BY vs view PK dependency

VENUS-200 venus Done high

Active-views refactor: DB views for deletedAt/isDemo filtering

VENUS-199 venus Done high

Futuro badge overcounted via participation join

VENUS-198 venus Done high

Fix recurring React #418 hydration mismatch on /practicas/[id] (ICU time-separator class)

VENUS-194 venus Done high

No-silent-failure contract: audit + enforce try/catch→appEvents+friendly toast, required-field validation toasts, swallowed-catch elimination, error→live-PM DM

PLUTO-529 pluto Done high

No-silent-failure contract (Mars lane): audit census + author shared mutation-wrapper/form-hook/lint-guard + codify contract in commons

MARS-408 mars Done high

Hub: brainstorm-brief to an aro with an open session silently DISCARDS the brief body (hub.mjs:2585) — make deflect non-destructive

MSG-261 llmmsg-srv Backlog high

Pluto IDOR: createNotification + markComentarioNotificationsRead unauthenticated, client-forgeable userId

PLUTO-530 pluto Done high

Operador may create práctica with 'Sin asistente' (null asistente); null-safe all KPIs/queries/reports/filters

PLUTO-528 pluto Done high

Comisiones grid UX: row-click opens existing detail page + relocate rename + hide delete unless empty (A+B)

PLUTO-525 pluto Done high

No-roles real-time aro:pluto alert channel (part 3 of 518)

PLUTO-519 pluto Done high

Assign active docente_jtp role to elazar+jtp test persona

PLUTO-521 pluto Done high

No-roles login denial: instrument silent page.tsx path + raise ARO alert

PLUTO-518 pluto Done high

Scaffold bare Next.js app + push first commit to empty GitHub repo

ENAMEL-5 enamel Done high

Legal review: HC data residency under Ley 25.326

ENAMEL-1 enamel Backlog high

/informes '#51 Realizada por docente' card is a dead metric — always 0/0% (queries dificultadPercibida for a label that can't exist)

PLUTO-513 pluto Done high

/informes 'En esta página' TOC card: rename to 'Secciones en esta página' + fix title/description run-together wrapping bug

PLUTO-512 pluto Done high

Entity cross-linking sweep: make named practica/comision/teacher/student references clickable, discoverable links

PLUTO-493 pluto Done high

Runtime ReferenceError: SupportActionResult is not defined at /practicas (system/serverError)

PLUTO-491 pluto Done high

Tidy sort core for portability + mark canonical SSOT (fleet sortable-table primitive)

PLUTO-484 pluto Done high

Harden /api/log-error (auth+rate-limit+body-cap)

MARS-393 mars Done high

Codex audit — P0 security/forensics: resume-link fail-closed + appEvents storage-key strip (7 sites) + userRoles role-change changelog

MARS-38 mars Done high

Harden INNER JOINs to lookupOptions that drop a live row if a status/programa lookup is ever deleted (soporte + alumnos)

MARS-166 mars Done high

/informes fn_getInformesBundle 25s exec + pg-pool connection exhaustion (469 perf regression)

PLUTO-471 pluto Done high

Mars DB scaling: migrate serverless runtime from direct-5432 to Supavisor transaction pooler (6543); reserve direct-5432 for weekly-kpi cron + migrations/seed; cut runtime pg.Pool max 16->~3; audit+cap per-request concurrent-query fan-outs (mirror Pluto mapWithConcurrency); pre-flight named-prepared-statement check (txn mode drops them). 3-PM consensus bs-mql6kqmrm6j, Venus=reference (already on pooler). One-env-var + cron carve-out, reversible.

MARS-145 mars Done high

Stage solo-ayudarg agent: config diet + auditsolo one-shot + Class-A push-gate + sibling clone + knowledge-file

AYUD-112 ayudarg Done high

Server-error email from instrumentation.ts onRequestError never delivers (8x SMTP ETIMEDOUT since 2026-06-10; SMTP healthy, instrumentation-hook teardown timing)

PLUTO-161 pluto Done high

CR 33: 10 items de Diagnóstico clínico radiográfico (obligatorios): 1) Enfermedad periodontal avanzada; 2) Fractura dentaria vertical/horizontal; 3

PLUTO-259 pluto Done high

Agent name parsing: forbid dashes inside role/project (single-token parts only)

OPS-72 agent-ops Backlog high

getSolicitudAnalysisContext: identity-match false negative — claimed 'no users row' for solicitud aa4ff295 despite pre-existing user veronica1994rios@gmail.com

PLUTO-385 pluto Done high

Post-/compact re-register unreliable for some cc agents (intermittent) - bootstrap silent-exit on hub blip + shim path + .agent-name fallback

MSG-223 llmmsg-srv Done high

llmmsg-srv team restructure: PM + support agent + on-demand coders; rename llmmsg-srv-cc->hub-llmmsg-srv-cc, retire llmmsg-srv-ca (spec+plan in docs/superpowers/)

MSG-130 llmmsg-srv Done high

Unify JTP and Ayudante: identical roles app-wide (caps + UI)

MARS-281 mars Done high

Admin: assign roles UI missing

AYUD-80 ayudarg Done high

share global ~/.claude rules to venus via /gdrive/temp/

MARS-280 mars Done high

Comunas en /necesidades: no se seleccionan de a 1

AYUD-19 ayudarg Blocked high

eq (Elazar Questions) — decoupled-from-bwi pending-Qs tracker; 3-PM consensus + Elazar approval pm-mars-cc-18608/18632; venus stub in flight

MARS-289 mars Done high

Memory-discipline policy: make enforcement discoverable fleet-wide (whey/venus/lezama)

MSG-179 llmmsg-srv Backlog high

└─ Land canonical memory-discipline.md in sh.git (docs/md), commit+push; nw-* pull propagates to 3 hosts

MSG-181 llmmsg-srv Backlog high

└─ Consolidate evolutiva-memory-discipline.md into fleet-neutral canonical + scrub project/agent names for public sh.git

MSG-180 llmmsg-srv Backlog high

└─ Per-host global CLAUDE.md: add memory-discipline.md to Lazy-loaded + 2-line enforcement pointer (whey/venus/lezama, one commit each)

MSG-183 llmmsg-srv Backlog high

└─ Stub evolutiva-memory-discipline.md -> sh.git canonical; update maintainer-agents.md pointer

MSG-182 llmmsg-srv Backlog

└─ Acceptance gate: fresh agent on each host must cite the policy+thresholds when asked is-there-a-md-policy

MSG-184 llmmsg-srv Backlog

└─ memory-discipline.md threshold divergence: prose (§10/§13/§15 ~4k tok/150-200 lines/8000 budget) != enforced memlint-cron.sh (600 warn/1200 trim); make doc reference script as threshold SSOT, drop duplicated numbers

MSG-185 llmmsg-srv Backlog

Admin UI notification for new /sumate solicitudes (currently only email to Belén)

AYUD-79 ayudarg Done high

Students may edit submitted prácticas until docente clicks Aprobar/Desaprobar

MARS-282 mars Done high

Email migration notice: 31 non-standard-email users

MARS-275 mars Done high

Agent-hygiene: automated CLAUDE.md/memory bloat detection + LLM compression (fleet-wide)

MSG-160 llmmsg-srv Backlog high

└─ memlint-runner CLI (sh.git): scan/trim CLAUDE.md+MEMORY.md with token counts, banned-phrase rules

MSG-161 llmmsg-srv Done high

└─ claude-md-trim.sh + trim-cwd-claude-md skill: LLM-driven compression wrapper (git-checkpoint before rewrite)

MSG-162 llmmsg-srv Done high

└─ memlint-sweep daily timer (venus-side): same as #whey, scoped to venus agents

MSG-164 llmmsg-srv Done

└─ memlint-sweep daily timer (whey-side): scans all agent CLAUDE.md, DMs Elazar+owner if bloated

MSG-163 llmmsg-srv Done

└─ memlint-sweep daily timer (lezama-side): same as #whey, scoped to lezama agents

MSG-165 llmmsg-srv Done

└─ /agent_hygiene endpoint: aggregate token-size telemetry from per-host sweeps

MSG-166 llmmsg-srv Done low

└─ chat-duo Agent Hygiene pane: surfaces top-bloat agents from /agent_hygiene

MSG-167 llmmsg-srv Backlog low

Add 'Rectificar' approval state: docente-only, comment required, email + in-app notification to student

MARS-283 mars Done high

Solicitar información button: add operator message field + ensure send to user works

MARS-284 mars Done high

Provision access: María Miyauchi Harumi + Daiana Pilla (Comisión F)

MARS-285 mars Done high

agent-local context-warn statusLine hook (clone pluto reference at 100k/150k/200k/300k)

MARS-276 mars Done high

Backup pluto DB + delete practicas test data — testing phase ended, client wants test data wiped. Order: (1) pg_dump backup to /gdrive/backup/, (2) delete practicas test rows + dependent FKs, (3) report counts.

PLUTO-388 pluto Done high

Document email-send playbook + no-deferred-jobs rule (Elazar tag 11573 b+c)

MARS-274 mars Canceled high

evolutiva-commons: 'do not defer' rule (do now or open WI; never silent defer)

MARS-273 mars Done high

drop dangling icloud auth row eb5854c8 (Luana)

MARS-272 mars Done high

ship per-app /api/internal/send-email + drop deferred lockdown commons rules

MARS-271 mars Done high

count users with non-gmail/non-pensanta/non-institutional login emails + email all

MARS-270 mars Done high

mergeUsers: set loser.surrogateOfUserId=keeper at soft-delete (audit BLOCK)

MARS-269 mars Done high

Luana Lazarte: update existing user email+authId to gmail, close pending accessRequest

MARS-266 mars Done high

merge users Felicitas Cava: A=81694c6a (old, dangling auth) → B=0872484c (active Google OAuth)

MARS-265 mars Done high

Lock sqlite mirror scope to lookupOptions; drop PII; gitpush-pre mtime lint

MARS-264 mars Done high

Fix /administrar/docentes/[id] RSC crash UNAUTHORIZED

MARS-262 mars Canceled high

Investigate Supabase MCP permission errors when creating views — db-mars-cc fell back to psql (db-mars-cc-20437). Likely missing GRANT for MCP role. Affects all 3 projects.

MARS-291 mars Done high

Force-clear stale Supabase auth cookies on next visit (one-shot reset)

AYUD-83 ayudarg Done high

Hide email+password login UI (Google-only), keep code path intact for easy revert

AYUD-82 ayudarg Done high

fix /administrar/docentes/[id] RSC crash (digest 1647002915, eventId c3cd6e9c)

MARS-261 mars Done high

Admin: delete data (different from self-erasure CR-290)

AYUD-81 ayudarg Backlog high

P1: Vercel git integration silent — 22h+ backlog, no auto-deploy since b18312b (2026-05-15-0258)

PLUTO-411 pluto Done high

Soft-land all RSC errors; trace digest 4277881123

MARS-260 mars Done high

audit codebase: wrap server components/actions/handlers in try/catch + reusable soft-landing error UI; trace digest 4277881123 (RSC sanitized prod error)

MARS-259 mars Canceled high

/notificaciones page has dead/old paths leading to 404 — audit links, fix or remove the page

MARS-258 mars Done high

Welcome email on user creation: createAlumnoWithUser (and other create-user surfaces) sends Spanish welcome email via existing nodemailer mailer.ts; fire-and-forget, graceful no-op if SMTP not configured

MARS-257 mars Done high

P0 BUG: /practicas/<uuid> 500 crash on detail view — broken for all approvers. researcher report agents/researcher-pluto/reports/ux-audit-2026-05-16.md item 1.

PLUTO-416 pluto Done high

Bedel can create alumno from /administrar/alumnos: full form on-page (pulldowns), no popup, no two-step user-then-role

MARS-256 mars Done high

Lane B: /administrar/usuarios/[id]/editar — unified user-edit page mirroring crear-form (kind picker + identity + roles + comisión + JTP + ayudante + periodo); replaces modal-in-grid + read-only detail pages

MARS-253 mars Done high

Update /sumate success copy: 'Gracias por tu interés en sumarte. Te contactaremos pronto en relación a tu solicitud. ¡Hasta pronto!' (remove old 'Datos enviados' header + old text)

AYUD-73 ayudarg Done high

Lane A: /practicas pagination + filters + drop HARD_LIMIT — wire existing filter-bar/pagination/search-input components, URL-synced searchParams, default pageSize=50, server-pushed filters

MARS-249 mars Done high

ARO topology v3 fleet-wide rollout (epic)

MSG-174 llmmsg-srv Backlog high

└─ Phase 2: hub aro_join validation + PM-election fallback + aro_list role surface

MSG-175 llmmsg-srv Done high

└─ Phase 3: launcher .aro v2 parser (ccs.sh/ccs-tmux.sh/bootstrap)

MSG-176 llmmsg-srv Done high

└─ Phase 5: chat-duo + chat-duo-web tier-aware UI

MSG-177 llmmsg-srv Done

└─ Phase 6: llmmsg-protocol.md §ARO Membership + §Routing rewrite

MSG-178 llmmsg-srv Backlog

Calles: Capital streets missing in picker (e.g. Av. Rivadavia not found)

AYUD-77 ayudarg Done high

Missing 'Yo me ocupo' button somewhere — client report

AYUD-78 ayudarg Done high

Add app version badge to /sumate page

AYUD-71 ayudarg Done high

ROTATE leaked secrets: DATABASE_URL, SUPABASE_SECRET_KEY, AUTH_GOOGLE_SECRET were committed to git history in .env.production (now deleted but historical commits still expose them)

AYUD-70 ayudarg Canceled high

Remove real .env.local and .env.production from disk (secrets-at-rest); replace with .env.local.sample placeholder per commons

AYUD-69 ayudarg Done high

UX nav: mover Admin al pulldown; Panel pill siempre visible en barra superior

AYUD-86 ayudarg Done high

Mobile UX + best-practices enhancement of pensanta.com (keep close to com.ar, no bugs or bad code)

HOAC-14 ha In progress high

└─ SEO + structured data + canonical URL setup (canonical = https://hormigonacero.com.ar/ pre-cutover)

HOAC-15 ha In progress high

└─ image optimization: CLS + webp

HOAC-16 ha Done

Support ticket #1: Maria Azul Etcheverry cannot submit practica at /practicas/nueva - reports 'filled all required data + radiographs but it will not let me send it' (2026-05-20 17:47 -03). Investigate, reproduce, fix+push if code bug, else explain cause + what is missing.

MARS-324 mars Done high

dbcaba backup gaps: dev backup targets wrong db + gdrive backup-mirror broke 2026-05-16 (dbcaba has no auto-backup today)

OPS-76 agent-ops Backlog high

Sumate: send Gmail SMTP from soporte-ayudarg@pensanta.com to belengalvan@catedraldelafe.org (replaces Resend path)

AYUD-68 ayudarg Done high

practica-form client-side required-field validation: on submit, if required fields empty, block submit, scroll to first missing field + show clear toast (no server round-trip)

MARS-326 mars Done high

Sidebar link directo a /admin/inactividad (Puia lo necesita visible, no enterrado en /admin dashboard)

PLUTO-355 pluto Done high

Support ticket #2: Jose Santiago Gonzalez has alumno profile but reports he is a docente - role misconfigured

MARS-327 mars Done high

Support ticket #3: Juana Frega gets an error when submitting a practica at /practicas/nueva

MARS-328 mars Done high

supportTickets close-and-notify: closing a ticket in /administrar/soporte emails the user the resolution text

MARS-329 mars Done high

ABM de prestaciones por entidad

AYUD-31 ayudarg Done high

global ~/.claude/CLAUDE.md HOT/COLD split to cut per-turn token load fleet-wide

EVO-44 evolutiva Done high

Chip UX fix: contrast/selected/unselected/category tints on needTag chips

AYUD-66 ayudarg Done high

gitpush.sh / post-push hook: verify Vercel deploy success before declaring done (3 commits silently failed type-check on prod)

AYUD-65 ayudarg Done high

KPI row collapses to column on mobile — should stay side-by-side (Total/Sin calificar/Aprobadas/Desaprobadas)

MARS-225 mars Done high

Cross-review WI 281+282+325 (recursos filters)

AYUD-63 ayudarg Done high

Backfill organizations.communeId — only 3/43 active orgs have it populated

AYUD-62 ayudarg Done high

RecursosView: replace hardcoded NEED_FILTERS with data-derived categories (same root cause as WI 280)

AYUD-61 ayudarg Done high

Cross-review WI 280 (mapa filtra por necesidad fix)

AYUD-60 ayudarg Done high

UI/UX cohesion audit — pre-client meeting deep-dive: per-role flows, surrogate-account removal, cross-surface comisión+jtp+ayudante consistency, IA consolidation

MARS-246 mars Done high

└─ Twin-route in-place UI unification (deferred from WI#213)

MARS-255 mars Done

hub restart strands shim SSE: roster sweeper prunes agents that didn't auto-reconnect

MSG-119 llmmsg-srv Done high

multi-subscriber SSE + sender echo + client dedupe (chat-duo + chat-duo-web cohabit)

MSG-118 llmmsg-srv Backlog high

lezama: add systemd restart guard for hub + fix autossh SIGTERM regression

MSG-114 llmmsg-srv Done high

PM seat lost on compact: pruneStale strips recorded-PM ARO membership

MSG-219 llmmsg-srv Done high

lezama watchdog: email notify-elazar.sh when zombie-fix fails

MSG-115 llmmsg-srv Done high

venus: whey-side watchdog for zombie detection (parallel to lezama watchdog)

MSG-116 llmmsg-srv Done high

tunnel down fallback alert: SSH + notify-elazar.sh when llmmsg unreachable

MSG-117 llmmsg-srv Done high

Security: getMatchCandidates server action missing requireAdmin() guard

VENUS-73 venus Done high

whey gdrive-bisync broken (aborting every ~30s, 'cannot find prior Path1/Path2 listings') since ~2026-05-16: evolutiva/venus/odaia backups local-only in /gdrive, not reaching Google Drive; needs rclone --resync recovery (path2/local-wins) on Elazar GO

OPS-77 agent-ops Done high

/administrar/usuarios/crear: dedicated form replacing modal/inline 'Agregar usuario' — kind picker (teacher/student/other) with contextual selects

MARS-243 mars Done high

ca.sh codex launcher: run codex on llmmsg-srv (ca= identity, sandbox-off, pull-only inbound, push-to-ccs) - fleet

MSG-186 llmmsg-srv In progress high

└─ Land reviewed ca.sh in sh.git (after PM draft review); commit+push

MSG-188 llmmsg-srv Backlog high

└─ Design + LIVE-TEST ca.sh: reads ca= from .agent-name, codex --dangerously-bypass-approvals-and-sandbox, injects -c mcp_servers.llmmsg-srv.env.LLMMSG_AGENT_NAME + LLMMSG_HUB_URL; PROVE shim auto-registers under ca name (roster check) before calling done

MSG-187 llmmsg-srv Backlog high

└─ Deploy/verify ca.sh on venus (sh.git pull); confirm a codex test session registers + hub URL resolves remote

MSG-189 llmmsg-srv Backlog

└─ lezama ca.sh deploy: GCABA box - get Elazar explicit OK before nw-lezama-cc syncs; codex AGENTS.md convention (aro_join on start, cm/cms -> read_unread)

MSG-190 llmmsg-srv Deferred

└─ Fix static ~/.codex/config.toml footgun: wrong key LLMMSG_AGENT->LLMMSG_AGENT_NAME + LLMMSG_SRV_HOST 127.0.0.1->public; bare codex (not via ca.sh) currently falls back to cc= in /opt/llmmsg-srv = registers as pm-llmmsgsrv-cc on loopback. Non-blocking; ca.sh -c overrides already bypass it

MSG-191 llmmsg-srv Backlog

bwi Tier 1 enhancements: add guard, mine banner, rot, status notify, evolutiva doc

OPS-38 agent-ops Done high

Cross-review WI 284+285 (logo responsive + ocupado→tomada rename)

AYUD-55 ayudarg Done high

lezama hub tunnel stale-bind restart thrash (llmmsg-lezama-tunnel.service, counter 3124)

MSG-193 llmmsg-srv Done high

.llmmsg-aros -> .aro consolidation: one per-agent ARO file (move aros[] into .aro, delete .llmmsg-aros, repoint resolve-aros.sh + bootstrap)

MSG-220 llmmsg-srv Done high

└─ ccs.sh: stop writing now-dead .llmmsg-aros + fix stale 'canonical' source comment (#631 phase-4; must precede phase-3 delete or files regenerate on launch)

MSG-224 llmmsg-srv Done

└─ .llmmsg-aros cleanup: remove 16 repo-tracked copies in pensanta website repos (#631 tail)

MSG-225 llmmsg-srv Backlog low

P-2: email-domain allowlist gate + deactivate 8 non-allowlist users (db-pluto-cc-30353/30429)

PLUTO-426 pluto Done high

P1: add wi doctor integrity checks

BSQ-3 basquet Done high

Eliminar datos (GDPR): opción de borrado por protección de datos

AYUD-30 ayudarg Done high

Asistente picker en /practicas/nueva no permite seleccionar — cliente reporta no poder cargar ayudantes/asistentes

PLUTO-222 pluto Done high

lezama split-brain: island-b hub (llmmsg-srv.service) auto-started, isolated lezama agents from whey hub

MSG-194 llmmsg-srv Backlog high

Conversación se actualice cada 24hs

AYUD-29 ayudarg Blocked high

gdrive bisync down since ~May 16 (whey<->venus) — venus agents loading 2-week-stale evolutiva shared .md rules (missing WI#450 + A+B + any ratified change since May 16); fleet correctness/drift risk. Fix = repair rclone bisync (maintainer territory), then re-measure per-agent preload vs true current files.

EVO-43 evolutiva Done high

Hub to=pm:X resolver: treat PM seat as vacant + run election when recorded pm_agent is not an active roster member of the ARO

MSG-131 llmmsg-srv Done high

cc-context-monitor: forced auto-/compact via bootstrap hook (over-budget agents ignore nudges)

MSG-195 llmmsg-srv Done high

backup-whey.service failing since 2026-05-09 02:03

OPS-31 agent-ops Canceled high

Hub prune-on-write deletes roster row -> cursor reset to 0 -> full-history message redelivery on reconnect

MSG-197 llmmsg-srv Done high

Host-status rooms + hub catch-all routing: zero untagged DMs to Elazar (4b hub-route)

MSG-222 llmmsg-srv Done high

chat-duo: remove Directory window/panel from sidebar (Elazar 2026-05-22-23:07)

MSG-134 llmmsg-srv Done high

Shim auto-register on startup from .agent-name; relaunch system-prompt covers -c continue (not just /clear/compact) - prevents post-relaunch poll-loop silence

MSG-135 llmmsg-srv Done high

P-3: assign 'bedelia' rol to lorena.salomone (asesora pedagógica per client; needs full-view + change comisión cap). per Elazar -31043

PLUTO-432 pluto Done high

backup-whey.service: NOPASSWD sudo for cryptsetup+mount in /etc/sudoers.d/backup-whey (failing since run #23 2026-05-08T14:22 - sudo prompts for rob pw under systemd no-tty); owner: os-whey-cc-w

OPS-32 agent-ops Backlog high

UX app-wide: diferenciar visualmente botones vs badges vs tarjetas

AYUD-85 ayudarg Backlog high

Global PreToolUse enforcement hook: deny non-llmmsg-srv-team agents Edit/Write to /opt/llmmsg-srv/{hub,bridge,scripts/lib}/** + deny Bash curl/http to hub privileged endpoints (extend existing PATH-script-block hook). Elazar approved ship - recurring: agents edit hub code/curl API instead of MCP tools.

OPS-79 agent-ops Backlog high

Perfil: mostrar/aclarar tipo de permiso del usuario

AYUD-28 ayudarg Done high

Cross-review WI 283 (new /necesidades/nueva create flow)

AYUD-56 ayudarg Done high

Columna 'bien' en /datos renombrar a 'necesidad'

AYUD-27 ayudarg Done high

Approval flow crea alumnos huérfanos: approveAccessRequest no inserta studentAssignments — orphan estudiantes no pueden cargar prácticas porque getClassmates retorna vacío. Fix: requerir comisionId+jtpId al aprobar (mismo patrón que /admin/usuarios manual create) e insertar studentAssignment en la misma tx. Backfill: 10 alumnos sin asignación en periodo actual.

PLUTO-345 pluto Done high

Comb entire app for raw UUID/ID shown in UI instead of label; fix all (reported on /informes JTP/Ayudante column)

MARS-335 mars Done high

evolutiva-backup down for mars+pluto: Supabase removed legacy db.<ref>.supabase.co direct hostnames (NXDOMAIN) - DSNs need pooler connection strings

OPS-67 agent-ops Done high

Reemplazar 'acción' por el nombre real de la acción en UI

AYUD-26 ayudarg Done high

Status 'ocupado': qué significa? aclarar o renombrar

AYUD-25 ayudarg Done high

Practicas list+detail page-ready views (P1 perf)

PLUTO-362 pluto Done high

Plugin hygiene v2 - rollout completion (lezama + deferred items)

OPS-58 agent-ops In progress high

└─ Plugin hygiene v2 - lezama host portion: telegram uninstall + host-global lean flip + movilba per-agent enables (see parent 459)

OPS-59 agent-ops Backlog high

Close aro_config auth gap: token-gate POST /aro_config + other privileged hub writes so a forged caller=elazar fails (coder-applog-cc exploited it to flip aro:applog PM). Elazar approved ship.

MSG-201 llmmsg-srv Backlog high

Alert when a CC session is blocked on interactive prompt (AskUserQuestion menu / permission approval / ExitPlanMode): wire CC Notification hook (or transcript-watcher fallback) to DM elazar via llmmsg-srv naming the blocked agent+host. Solves operator having to hunt tmux windows. Research dispatched to claude-code-guide.

MSG-202 llmmsg-srv Backlog high

Date Handling commons + fix periodos-grid.tsx + periodos/[id]/page.tsx TZ anti-pattern

MARS-287 mars Done high

lezama host-dark recurrence: root-cause + harden against full-dark (2nd occurrence; died 2026-06-01 18:03 ART, clean cliff)

OPS-80 agent-ops Backlog high

└─ Reboot forensics: journalctl -b -1 -k (panic trace=crash / empty=power loss) + full -b -1 (OOM?), dmesg|tail-50, df -h, last -x reboot, smartctl -a

OPS-81 agent-ops Backlog high

└─ Validate the installed recovery software (failed twice): identify unit, check prior-boot status+journal (fired/tried/failed?), confirm dead-kernel/power is below userland scope; validate reverse-tunnel health EXTERNALLY from whey, never trust the ssh_r_active self-report flag

OPS-82 agent-ops Backlog high

└─ Root-cause-gated remediation (after forensics): OOM->add swap + tune vm.overcommit_memory; disk-full->monitor cron + notify-elazar; kernel panic->enable kdump for post-mortem

OPS-85 agent-ops Blocked

└─ Scope a HARDWARE remote-reset path for dead-kernel recovery (autossh/userland watchdog cannot revive a dead kernel): IPMI/iDRAC if present, else smart PDU - only no-physical-access option

OPS-84 agent-ops Canceled

└─ ZeroTier transport redundancy: zt_whey_visible=0 across all 5435 probes -> single public-dns path. Bring ZT up (zerotier-cli listpeers, whey visible) as a real fallback route

OPS-83 agent-ops Canceled

└─ lezama single-path fragility: all 5435 probes since May23 are public-dns with zt_whey_visible=0 (ZeroTier fallback never worked); make ZT a real 2nd route + fix phantom ssh_r_active self-report

MSG-205 llmmsg-srv Backlog

└─ lezama host-dark alert: watchdog fires notify-elazar+eq after 3 consecutive 'unreachable' skips (~15min), instead of silently skipping forever

MSG-204 llmmsg-srv Backlog

└─ lezama out-of-band reset: a dead kernel cannot self-recover; scope IPMI/iDRAC vs smart-PDU for remote host-dark power-cycle (blocked on Elazar: does box have IPMI?)

MSG-206 llmmsg-srv Backlog low

viewAllPracticas cap re-ship: bedel role gets cross-comision /practicas read-only

MARS-288 mars Done high

Ayuda y Soporte universal feedback button

MARS-311 mars Canceled high

Logo deforme, no responsive

AYUD-24 ayudarg Done high

No permite cargar necesidades

AYUD-23 ayudarg Done high

Recursero: filtro por necesidad no anda

AYUD-22 ayudarg Done high

Necesidades: no se seleccionan de a 1 (multi-select roto)

AYUD-18 ayudarg Blocked high

Recursero: filtro por comuna no anda

AYUD-21 ayudarg Done high

Mapa no filtra por necesidad

AYUD-20 ayudarg Done high

Olvidé clave

AYUD-46 ayudarg Done high

Sumate no anda

AYUD-17 ayudarg Done high

nw-agent launcher template: haiku + --effort low + autoCompact + loop self-monitor rule

OPS-37 agent-ops Done high

Transport reliability + observability program (whey/venus/lezama)

MSG-149 llmmsg-srv Backlog high

└─ Idle over-budget agents never self-compact: wire host-side actuator (whey/lezama)

MSG-207 llmmsg-srv Backlog high

└─ channel-rule audit+install scripts (fleet rollout)

MSG-169 llmmsg-srv Done high

└─ TRANSPORT.md SSOT: per-host transport, fallback, watchdogs, runbook

MSG-150 llmmsg-srv Done high

└─ llmmsg-doctor CLI: one command, transport+latency+last-good+units

MSG-151 llmmsg-srv Done high

└─ host_probes table + bootstrap-probe (every session start posts transport snapshot)

MSG-152 llmmsg-srv Done high

└─ Fleet context-cost rollout: send-keys /compact watchdog + aro_leave-on-idle

MSG-210 llmmsg-srv Backlog high

└─ aro_leave-on-idle rule: llmmsg-protocol.md ARO-Membership canonical + per-agent CLAUDE.md one-liner (self-leave non-essential AROs when last task done, stay registered, rejoin on assignment)

MSG-212 llmmsg-srv Done high

└─ send-keys /compact watchdog in cc-context-monitor.sh: fire tmux send-keys -l /compact + C-m on cost-cap cross; pane=session_name==agent; skip mid-gitdeploy

MSG-211 llmmsg-srv Done high

└─ cc-context-monitor: fix stale over_cap_alert fallback DM (~L668) - stop instructing dead ScheduleWakeup /compact; send-keys is primary, fallback should inform/escalate not hand a dead path. Fold into next bump (with WI#614). Align messaging with hub #621.

MSG-215 llmmsg-srv Done

└─ cc-context-monitor safety-net: flag/extend nag for agents that miss self-leave (seq AFTER rule lands)

MSG-213 llmmsg-srv Backlog

└─ WI592 nag: drop CronCreate /compact from emit_context_budget_nag interactive branch once send-keys watchdog verified fleet-wide

MSG-214 llmmsg-srv Backlog

└─ cc-context-monitor measured role-cap version (floor+EOQ headroom) - successor to #617 seed caps

MSG-217 llmmsg-srv Backlog low

└─ path-B UserPromptSubmit injection copy: reframe CronCreate as automatic-fallback not manual action (avoid conflict w/ v2.13 over-cap DM)

MSG-218 llmmsg-srv Backlog low

└─ Lezama agent auto-bootstrap: cold-start agent online + AROs joined + hub-reachable with zero prompt-paste

MSG-159 llmmsg-srv Done high

└─ Whey-side silence detector implementation (claimed by nw-whey-cc)

MSG-154 llmmsg-srv Done

└─ chat-duo v1.19.0 Fleet Health pane (depends on /fleet_health endpoint)

MSG-155 llmmsg-srv Done

└─ /fleet_health endpoint + transport-registry table

MSG-156 llmmsg-srv Done

└─ Venus-side watchdog + alert standardization (mirror lezama pattern)

MSG-157 llmmsg-srv In progress

└─ Lezama-side daemon probe: llmmsg-lezama-daemon-probe.sh + systemd timer (mirror of whey #536)

MSG-168 llmmsg-srv Done

└─ fleet_health: inferHostFromAgent emits bogus host='w' from -cc-w suffixed agents

MSG-172 llmmsg-srv Done

└─ Fleet Health false-green: /fleet_health reports host green while host-dark (hub.mjs:868-878 probe-age tripwires too lax + agents:[] reads as healthy)

MSG-203 llmmsg-srv Done

└─ Hub-side silence detector: alert when host goes dark > threshold

MSG-153 llmmsg-srv Done

└─ chat-duo-web 9704 to lezama: move off whey-initiated -R (VPN-dependent) to lezama-initiated -L over public route

MSG-208 llmmsg-srv Backlog

└─ Shim transport prefer-list: loopback -> ZT -> public DNS (defer until W3+W4 prove design)

MSG-158 llmmsg-srv Backlog low

Pin hub registrations for cc-context-monitor-{whey,venus,lezama}

OPS-44 agent-ops Done high

Bug: Pendiente hace anchored to wrong date on practicas

MARS-220 mars Done high

UX /necesidades/(id): Estado+Prioridad en misma línea; botón Yo me ocupo debajo de Estado

AYUD-84 ayudarg Done high

UX /necesidades/(id): mostrar 'Tomada por [nombre]' junto al botón Liberar cuando estado=En Curso

AYUD-89 ayudarg Done high

memory-budget over-budget warning: route shared-file footprint breaches to project PM with escalation chain

OPS-57 agent-ops Backlog high

memory-budget Layer 1: gdrive-canonical 'Memory Budget' directive + footprint-check threshold (thread 3); incl. maintainer-agents.md memory-lint ownership section (thread 4)

OPS-48 agent-ops Done high

eq backend unbuilt: hub.mjs has no /eq_* routes + no elazar_questions table — eq CLI POSTs to non-existent endpoints (/eq_ls returns not_found). chat-duo Pending panel reads elazar_questions, so no Elazar question can ever surface. Implement hub-side /eq_add /eq_ls /eq_show /eq_answer /eq_dismiss /eq_stale + elazar_questions schema per eq-spec.md

OPS-49 agent-ops Backlog high

v2.7 over-cap remediation: parked over-cap agent still in its AROs gets woken into a full-context read-turn by every ARO fan-out (~30-60s) — park does NOT stop the burn. Fix: over-cap sequence becomes ss -> DM handoff to PM -> aro_leave ALL AROs -> idle. Edits: v2.6 over-cap DM text in cc-context-monitor.sh (bin-whey-cc) + wind-down line in global directive (nw-whey-cc). Fresh bin-whey-cc takes alongside thread 2

OPS-50 agent-ops Done high

chat-duo: collapse origin_aro-less agent->Elazar DMs into ONE shared room (stop per-agent room proliferation)

MSG-124 llmmsg-srv Canceled high

memory-lint §12/§9 vs global CLAUDE.md: per-agent identity + dates in memory/ — Elazar decision

OPS-52 agent-ops Done high

Trim auto-loaded context for mars/evolutiva sessions: evolutiva-commons.md (5.3k) + mars-commons.md (3.6k) + mars/CLAUDE.md (2.1k) + mars MEMORY.md (3.8k) = ~14.8k. Target <=8k total per-session auto-load. Editorial trim, Evolutiva PM executes. Detection side is WI 444 (memory-lint token alarm, 8k line). Surfaced by Elazar 2026-05-21 from a researcher-mars-cc /memory dump (~17k total).

EVO-42 evolutiva Done high

Cátedra branding sweep: APP_CATEDRA_NAME env + email subject prefix [Endodoncia FOUBA]

MARS-277 mars Done high

chat-duo-web tablet-grade evolution (path A for unified browser client)

MSG-171 llmmsg-srv Backlog high

Repoint chat-duo Elazar Pending panel to v2 eq source

MSG-123 llmmsg-srv Done high

llmmsg-srv roster & PM resilience: pruneStale soft-offline, offline buffering, PM re-election

OPS-54 agent-ops In progress high

└─ PM-liveness watchdog: cc-context-monitor detects a PM-less ARO, DMs Elazar

OPS-55 agent-ops Blocked high

└─ chat-duo: render origin_aro-tagged messages in their ARO room, stop spawning a room per DM

OPS-56 agent-ops Done high

└─ chat-duo side of always-a-PM (brainstorm items 1+7): drop client-side PM resolution + aro_config write, send to=pm:X, render resolved_pm in bubble, queued=true inline confirmation

MSG-126 llmmsg-srv Done

Audit quick HIGH wins — solicitudes N+1 (vAccessRequestsForReview LATERAL, 500q/load to 1) + analyzeModeOn=false (stop imageUploadAnalysisEvents +600/day growth) + practicas/[id] serial-await to Promise.all (mirror mi-cursada sibling).

MARS-41 mars Done high

KPI report generator (phase-1 read-only PG script)

BWI-4 basquetwi Done high

cc-context-monitor path-B/wind-down DM mis-targeting: db-mars-cc remediation delivered to nw-venus-cc

OPS-9 agent-ops Backlog high

Make context-budget nag idempotent: stop re-nagging every turn + stop stacking /compact crons after compact

MSG-12 llmmsg-srv Backlog high

Codex receive-bridge: sidecar tmux-poke so Codex agents auto-drain llmmsg unread (no live push)

MSG-10 llmmsg-srv Backlog high

└─ HARD PREREQ: derive + test a Codex-specific ready-prompt regex for idle-gating (CC matcher will NOT transfer; wrong matcher = corrupted-buffer poke). No send-keys goes live until validated.

MSG-11 llmmsg-srv Backlog high

Ship wi-web browsable UI on bwi.pensanta.com:8765 (HTTP-mode -> wi-http -> PG, systemd Restart=always)

BWI-3 basquetwi Done high

one-db: 5 telemetry sqlite -> Postgres17 consolidation (umbrella)

ONEDB-1 one-db Backlog high

└─ No dual-write: cutover = flip reporter (netreport.sh) to pg + drop sqlite writes after parity. Per Elazar 2026-06-06.

ONEDB-2 one-db Done high

└─ GATE#1 step iv: drop sqlite writes in 6 collectors after readers repointed+proven — pg-primary, JSONL-spool = resilience cache (Elazar: NO sqlite write-ahead). cc-context-monitor keeps LOCAL rolling cache for its own over-cap loop + pg mirror

ONEDB-11 one-db Done high

└─ GATE#1 step iii: flip venus-kpi-collect.sh to pg-write (telemetry_rw via metrics_write) — kills kpi.sqlite split-brain co-writer; nw lane (cross-host)

ONEDB-10 one-db Blocked high

└─ GATE#1 step v: remote-monitor ingestion — venus/lezama write whey-pg DIRECTLY per-host (NO ssh-pull). Transport: venus=ZeroTier 172.27.178.27 or Tailscale; lezama=ssh -w L3 tunnel only (GCABA blocks ZT+TS), pg client→172.30.0.1. whey: listen_addresses + pg_hba for both + telemetry_rw role. nw co-owns

ONEDB-12 one-db Backlog high

└─ GATE#1 step ii: repoint 5 sqlite readers to pg telemetry_ro — netreport.sh, testnetwork --report, ccusage-daily-kpi.sh, llmmsg-kpi-report.sh, memory-lint.sh

ONEDB-9 one-db Backlog high

└─ GATE#1 step i: add reporter-set pg views to schema.sql (network v_history/v_report/v_by_wifi; kpi 3-day window) — distinct from coder's dashboard views; back the 5 sqlite readers

ONEDB-8 one-db Backlog high

└─ DECISION (Elazar): make ~/Documents/terminal/one-db its own git repo? bin (PM) owns the call

ONEDB-7 one-db Backlog high

└─ SELECT * views: one wide view per table (+rollups) so agents never guess columns - default SELECT *, not named fields.

ONEDB-4 one-db Done

└─ kpi.pensanta.com dashboard on whey (chat-duo-web model): single-file Node .mjs reads telemetry views -> static KPI pages.

ONEDB-5 one-db Todo

└─ DECISION (Elazar): kpi.pensanta.com exposure - Tailscale-only (team rec: bind whey tailnet IP, no TLS/cert ops, no public attack surface) vs public-TLS

ONEDB-6 one-db Backlog high

└─ kpi-web MANIFEST: add Track A views (usage.v_kpi_tokens_7d + context.v_kpi_compact_7d) to server.mjs allowlist + restart

ONEDB-14 one-db Done

└─ Telemetry collector registry: lazy-load .md - per source list systemd unit/timer -> collector script -> what data -> which reporter DMs Elazar+aro.

ONEDB-3 one-db Done

one-db: consolidate all whey telemetry/KPI sqlite into one local Postgres 17 db (telemetry), schema-per-domain, SELECT*-friendly views, reporting web UI. Goal: one place for all KPIs/perf/context data + Elazar can actually find/read them.

OPS-2 agent-ops Canceled high

└─ one-db: cutover (NO dual-write — Elazar dropped it 2026-06-06). Collectors write straight to pg via metrics_write (spools to JSONL on pg-down, so no data loss); backfill already loaded full history. Flip 5 collectors + netreport off sqlite, retire sqlite files once pg-proven.

OPS-6 agent-ops Canceled high

└─ one-db: views layer — every table gets enough pg views (latest / daily / report rollups) so agents query a view with SELECT * (never name fields). Goal: zero token waste guessing columns. Document view list in telemetry-inventory.md.

OPS-4 agent-ops Canceled high

└─ one-db: lazy-load telemetry-inventory.md — single doc agents read on demand listing, per data stream: which systemd unit/timer/cron collects it, which script the unit runs, which table it writes, and which script reports to Elazar + which aro. Keep it the SSOT map of the whole telemetry pipeline.

OPS-3 agent-ops Canceled high

└─ one-db: kpi.pensanta.com web UI on whey — small read-only server (chat-duo-web model, single-file .mjs) that reads the SELECT* views and renders KPI pages, so Elazar has one habit-friendly place to glance at all KPIs and drill into any metric when something looks off.

OPS-5 agent-ops Canceled

agentteamlaunch.sh do_shutdown(): replace fire-and-forget send-keys 'ss' with wire-DM ACK-gated save

MSG-6 llmmsg-srv Done high

Ayudante grading-responsibility = NONE (Elazar policy B): ayudantes must NOT count toward sin-calificar NOR receive grading nags. getJtpStats pending/sin-calificar axis -> JTP (assignedTeacherUserId) ONLY (drop ayudante CROSS JOIN from that axis); weekly-kpi grading nags -> JTP only; drill auto-follows getJtpStats. Acceptance: Belén Montoya bcd98024 (pure ayudante) shows 0% sin calificar; no ayudante gets a grading nag. Fixes both Belén-100% + Florencia-nag complaints.

MARS-23 mars Done high

└─ Clean mislabeled appEvents: remove form-validation toast logger (toast/error-shown, level=error noise) + recategorize sa-teachers-cascade category admin→system

MARS-27 mars Done high

└─ Re-add ayudante notify at práctica creation as INFORMATIONAL (new type practicaAyudanteAsignada, in-app only, no email); JTP keeps grading practicaToReview+email. Elazar follow-up to MARS-23/25 — notify whoever was PICKED as JTP/ayudante, nobody else; no grading semantics for ayudante.

MARS-26 mars Done high

└─ Notifier de-nag (MARS-23 follow-up): notifyPracticaCreated (practicas/nueva/actions.ts:960-968) sends in-app 'Nueva práctica para revisar' (type practicaToReview) to BOTH JTP and ayudante via graderIds=dedup([jtpId,ayudanteId]). Drop ayudanteRecipientId -> notify JTP only. This is Florencia's original nag path (in-app); email already JTP-only. Correct under policy B (ayudante not a grader).

MARS-25 mars Done high

└─ weekly-kpi digest: stop nagging ayudantes for ungraded practicas (JTP-only pending per MARS-23 Elazar ruling) — REQUIRED before Monday 06-08 cron

MARS-24 mars Done high

└─ Block JTP unassign-to-null in /alumnos dropdown: remove the empty '-' option for the JTP (assignedTeacherUserId) column so a student's JTP can only be reassigned to another teacher, never cleared to none. KEEP '-' for the ayudante column (ayudante is optional). Add server-side guard in reassignTeacher rejecting null when field=assignedTeacherUserId.

MARS-29 mars Done

└─ /practicas: add visible teacher filter scoped to JTP-role only — show prácticas where the selected teacher is the JTP (assignedTeacherUserId, live), exclude prácticas where they're the ayudante. Surface the MARS-23 JTP-only ?jtp= drill as a first-class filter control.

MARS-28 mars Done

JTP/Ayudante live-vs-snapshot: reporting/drill/notifications must read CURRENT assignment not creation snapshot; null same-person backfill (ayudante=jtp); add both-roles guard

MARS-21 mars Done high

└─ student-assignment-tcm.ts peer-inference fallback (L188-234) seeds a NULL SA from OTHER students' practicas.jtpUserId/ayudanteUserId snapshot; TCM-gated (deletedAt IS NULL + role-matched) so stale teachers rejected — NOT a rule-(b) violation, low risk. Tighten to prefer live peer SA over práctica snapshot.

MARS-22 mars Backlog low

Post-VENUSINF systemd-client bearer sweep: applog-pull/listen + errscan inject LLMMSG_HUB_BEARER (hubPost hardcoded bearerless, applog-pull.ts:121) + per-instance bearer env across hosts - restores fleet alerting (applog-pull@pluto unauthorized live)

MSG-42 llmmsg-srv Done high

EVO-12 regression: captureServer404 in not-found boundary fires on EVERY page view (false page_not_found rows fleet-wide)

EVO-19 evolutiva Done high

└─ mars EVO-19 leg: two-sink fix; by-route 20x/login=1 floor likely higher under UA-tag; ship on body=(a) confirm

EVO-23 evolutiva Done high

└─ venus EVO-19 leg: two-sink fix (pull capture out of not-found boundary); corrected probe shows 200 GET /login wrote action=404/level=error/alert=true row firing real alert+email; demonstrated pollution -> ship

EVO-21 evolutiva Done high

└─ venus classifier-parity follow-up: add referer-discriminator to tierForNotFound so genuine-bad-routes tier stricter (ayudarg already has it); venus currently OVER-escalates - phantoms classified error/alert=true. OPTIONAL low-pri polish, NOT a fix: two-sink EVO-21 already removes the phantom source so this only affects FUTURE real bad-route tiering. Defer per pm-venus scope decision (do not fold into EVO-21 push).

VENUS-34 venus Done someday

└─ pluto EVO-19 leg: two-sink fix; UA-tagged causal proof (5x /login -> 5 phantom rows on 200) + matcher-excluded empty-route blindspot; ship

EVO-22 evolutiva Backlog high

└─ ayudarg EVO-19 leg: pull captureServer404 out of not-found boundary; two-sink (catch-all + [id] throw-site capture); verify valid routes stop writing page_not_found

EVO-20 evolutiva Done high

Post-/compact idle-agent wake: shim re-emit + cursor-on-drain (fix A+B)

MSG-4 llmmsg-srv Done high

Bug: PM kind=brainstorm in active ARO opens sibling session instead of being rejected

MSG-71 llmmsg-srv Done high

soporte/actions.ts: (1) close+notify race can double-email a user — make close conditional UPDATE…RETURNING, email only the transitioned row; (2) support update returns ok when zero ticket rows updated — treat empty RETURNING as not-found. Codex audit, Elazar-greenlit 2026-06-30. Investigate-then-fix; audit cross-review (notify/email path).

MARS-365 mars Done high

Option A: drop domain gate at login callback + allow non-allowlisted domains on access-request/approval path (Elazar decision 2026-07-02)

PLUTO-465 pluto Done high

hub: enforce ARO membership on lane-tag (1a) + fan-out (2a)

MSG-258 llmmsg-srv Done high

Rectificar-reminder mailer: bearer-gated /api/internal/rectificar-reminder, per-student email listing their rectificar prácticas as /mi-cursada/[id] links, sample/live-gated (confirm-token + cohort===39), live routes through emailOutbox (454-self-heal). Elazar directive 2026-06-30.

MARS-366 mars Done high

Soporte parity P2: server actions postSupportReply(staff) + closeAndNotify — reuse P0/PLUTO-203 shared email leg (sendAdminEmail, verified soporte-evolutiva-pluto@ From, idempotency, rate-limit); KEEP manageSupport-EXCLUSIVE gate (no fullAccess/manageUsers/adminConsole escape). Audit design-ping on email+gate before impl.

PLUTO-218 pluto Done high

Floating 'Ayudar' button overlaps/covers the pagination 'next' button on /practicas. Reposition (move up / adjust offset or z-order) so it no longer obscures the paginator controls. Elazar 2026-06-30. Reversible UI.

MARS-364 mars Done high

404-capture path exhausts Postgres connection pool during scanner floods (EMAXCONN)

MARS-346 mars Done high

Informes jefe-filter narrowing: KPIs/practicas count + roster must narrow to selected JTP (fn_getInformesBundle filterJtpIds param + page wiring) — Elazar video feedback 2026-07-02

PLUTO-467 pluto Done high

'Pendiente hace' column sorts as string not numeric (e.g. 10 days ordered before 2). Sort by the underlying numeric/duration value, not the formatted label. Likely /practicas. Elazar 2026-06-30. Reversible.

MARS-363 mars Done high

EPIC: agents-efficiency-KPI - per-agent context/token efficiency report Elazar actually sees (Elazar 2026-07-02: 'never seen a report on that'). Inputs: usage polls + context samples (venus PG). Deliverable: recurring per-agent report (tokens/turn, cache-read overhead, context growth, compact frequency) surfaced to Elazar. Coordinate: bin-whey-cc (one-db data), coder-llmmsgsrv-cc (kpi views/report script)

SAVETOKE-33 saveTokenz Backlog high

└─ one-db: retire non-essential telemetry fleet-wide per Elazar - drop network/dns/ping probes + dbsize + kpi.checks collection AND all cross-host telemetry shipping (lezama+whey write nowhere remote; supersedes ONEDB-15 repoint branch); KEEP usage polls + context samples on venus only (efficiency-KPI inputs + compact-monitor actuator data, pending Elazar veto on context)

SAVETOKE-34 saveTokenz In progress high

CR 7: Renombrar a Complicaciones Intraoperatorias

PLUTO-233 pluto Done high

/informes JTP-Ayudante aggregate contradicts drilldown: shows 0% graded/pending for a teacher who is up-to-date (Florencia Neput, uuid bcd98024); wrong grading nags sent. Suspect getJtpStats SA-gated graded axis (MARS-5/inert-oracle root cause)

MARS-12 mars Done high

Hub command /audit_cli: Elazar-only, sends channel-discipline self-audit prompt to last-messaged agent in the ARO

MSG-229 llmmsg-srv Done high

Epic: General — cross-fleet token reduction (message size, batching, routing)

SAVETOKE-1 saveTokenz Backlog high

└─ Fleet: never-multi-recipient audit — audit-mars-cc (394 msgs all 1-to-1), audit-mba-l-cc, db-basquetwi-cc never use array sends; coach + verify adoption

SAVETOKE-24 saveTokenz Backlog

└─ Fleet: pm-batching directive — PMs batch routine updates (FINISHED notices, progress) to Elazar; urgent (error/blocker/question) still immediate

SAVETOKE-25 saveTokenz Canceled

└─ Role/agent-scope the eager MEMORY.md index — needs ccs.sh launcher change. MEMORY.md index is the ONE memory file eager-loaded every turn; today it's keyed per-PROJECT (by cwd), so e.g. db-mars-cc's eager index carries coder-only entries it never uses. Re-key the index per-AGENT so each agent eager-loads only its own index lines (individual fact files stay lazy as today). Launcher/sh.git change (ccs.sh memory-path derivation). Owner: bin-whey-cc + pm-llmmsgsrv-cc coordinate (sh.git). Resolves Elazar's per-agent-memory premise at the eager layer where it actually costs tokens.

SAVETOKE-31 saveTokenz Backlog

└─ Codification-arrow discipline: graduate stable per-agent memory facts into ROLE-scoped commons, prune the memory copy. evolutiva-memory-discipline.md §4 already defines memory->commons graduation; make it routine so durable role-shared knowledge lives once in role-commons (shared, deduped) instead of being re-derived in each agent's memory. Pairs with the two structural WIs above; ongoing per-PM hygiene, not a one-shot. Cross-fleet discipline.

SAVETOKE-32 saveTokenz Backlog

└─ Role-scope the eager commons stack — biggest per-turn lever. Today every agent eager-loads the FULL CLAUDE.md chain each turn (global + project-commons + evolutiva-commons + pm-commons + all rules symlinks); memory-budget hook already flags ~13K>8K footprint. Split commons so each ROLE loads only what it needs (db agent doesn't carry coder UI/shadcn rules; coder doesn't carry db-hygiene/DSN rules). PM-mediated brainstorm (Pluto, 2026-06-27) reached unanimous PM convergence this is the real lever, NOT per-agent memory (fact files are already lazy-recalled, not eager). Owner: bin-whey-cc (commons are whey-canonical). Cross-fleet: affects all 3 apps.

SAVETOKE-30 saveTokenz Done

Epic: Venus — venus ARO token reduction

SAVETOKE-2 saveTokenz Backlog high

└─ PM→Elazar DM batching: only decisions+blockers go direct-DM; route routine status (FINISHED/deploy/progress) to ARO or 1x daily digest. Baseline pm-venus 67 DMs/44k chars to Elazar in 24h (highest char-volume of any PM). Target -60%.

SAVETOKE-14 saveTokenz Canceled

└─ pm-venus-cc send consolidation: collapse sequential status/decision within a 30-min window into one post; eliminate ack-only/phatic sends. Baseline 185 deliveries/50-per-hr peak/16 rapid Elazar exchanges. Target 185->140 deliveries.

SAVETOKE-15 saveTokenz Backlog

└─ Worker coordination OFF aro:venus: intra-team coder<->audit back-and-forth via DM; reserve ARO for PM summary + brainstorm turns only. Baseline aro:venus = fleet top-cost ARO, 77k tokens/416 msgs/24h.

SAVETOKE-16 saveTokenz Backlog

└─ coder+audit-venus large-ARO trim: design-justifications/audit findings post HEADLINE+3-bullet to ARO, full detail DM to PM/counterpart. Baseline coder avg 888 chars, audit 1100+ chars/msg; design-ping 4281-char samples.

SAVETOKE-17 saveTokenz Backlog

└─ audit-venus checkpoint batching: 3+ audit checkpoints within 2hr collapse to one batched post; PASS/BLOCK verdicts stay single-line. Baseline audit-venus 90 deliveries/24h, burst-flagged.

SAVETOKE-18 saveTokenz Backlog

Epic: Mars — mars ARO token reduction

SAVETOKE-3 saveTokenz Backlog high

└─ db-mars-cc tiny-send batching: 14 sub-52-char status acks/24h; fold working:/done: into fewer messages per task

SAVETOKE-13 saveTokenz Backlog

└─ audit-mars-cc: adopt array multi-DM (to:[...]) for shared findings — 394 msgs/24h ALL 1-to-1, same finding sent separately to pm+coder; ~10,600 tokens/day (biggest single Mars lever)

SAVETOKE-9 saveTokenz Backlog

└─ pm-mars-cc <-> db-mars-cc ping-pong: 27 rapid exchanges/24h (~6,315 tokens); batch dispatch+result into fewer round-trips, drop sub-30s over-checking

SAVETOKE-11 saveTokenz Backlog

└─ coder-mars-cc/audit-mars-cc oversized ARO design-pings: 4400-char 'never-directive' mandates to ARO; headline+summary in ARO, full detail DM-to-PM/assignee

SAVETOKE-12 saveTokenz Backlog

└─ pm-mars-cc: Elazar batch-digest protocol — 87 DMs/24h (fleet-highest count); route only decisions/blockers direct, fold FINISHED+status into 1 EOD digest DM

SAVETOKE-10 saveTokenz Canceled

Epic: Pluto — pluto ARO token reduction

SAVETOKE-4 saveTokenz Backlog high

└─ aro:pluto worker post-size: coder-pluto-cc (avg 1137 chars/delivery) & audit-pluto-cc (avg 1118) post full design-pings/audit-findings to the room. Move to headline+summary in ARO, full detail (file paths, gate criteria, evidence) DM to PM/assignee. Target ~400 chars/ARO post. Saves ~700 chars x ~150 posts/day.

SAVETOKE-27 saveTokenz Backlog

└─ aro:pluto reframe worker->PM mandate posts (KPI Finding 8): coder-pluto-cc 'never-directive' (4590 chars) + audit-pluto-cc 'do-not' (4465 chars) one-way mandates to pm-pluto-cc. Pattern: worker PROPOSES (concise) in ARO/DM, PM decides+announces. Removes 4k-char one-way mandate posts + clarifies accountability.

SAVETOKE-28 saveTokenz Backlog

└─ aro:pluto tiny-send batching (low ROI): db-pluto-cc (+ others) emit separate 'working:'/'done:' status acks (31-247 chars). Batch into single first+last-action message per task. Minor per-send tag/overhead savings; flag-only, not urgent.

SAVETOKE-29 saveTokenz Backlog

└─ aro:pluto PM->Elazar digest batching: route routine status/FINISHED/deploy-notices into 1 daily digest DM; only decisions+blockers go direct-DM; TLDR-first, cap <1500 chars. CARVE-OUT: triage-loop pre-analyzed verdicts + live-ticket coordination stay direct (Elazar's model wants them pushed). Highest ROI in Pluto area: pm-pluto-cc sent 39 DMs/35807 chars/~8952 tok to Elazar in 24h, highest avg chars/DM (918).

SAVETOKE-26 saveTokenz Done

Epic: llmmsg-srv — hub/MCP/infra token reduction

SAVETOKE-5 saveTokenz Backlog high

└─ Fix v_large_aro_content_24h view: add recipient LIKE 'aro:%' filter (currently over-counts DMs with origin_aro carry-through as ARO fan-outs — inflates coaching false positives)

SAVETOKE-19 saveTokenz Backlog

└─ Hub: implement /breaker auto-reset after N hours if tripped by legitimate traffic (today required manual reset, blocking Elazar DMs)

SAVETOKE-20 saveTokenz Backlog

└─ Haiku KPI: add fixed v_large_aro_content_24h + re-run coaching after view fix to get accurate per-agent attribution

SAVETOKE-21 saveTokenz Backlog

Epic: Movilba — mba-l ARO token reduction

SAVETOKE-6 saveTokenz Backlog high

Epic: NightWatch — night-watch processes token reduction

SAVETOKE-7 saveTokenz Backlog high

Epic: AlertsAndStatusMessages — alerts, status, scraper messages token reduction

SAVETOKE-8 saveTokenz Backlog high

└─ context_monitor scraper: 12 msgs/day × 966 chars each = 11.6k chars to agents already over budget — add backoff (stop sending after 3 consecutive over-budget pings; resume only after /compact)

SAVETOKE-22 saveTokenz Backlog

└─ KPI scraper: fan-out to 5 recipients on each send — switch to single DM to pm-llmmsgsrv-cc only; PM decides if fleet needs it

SAVETOKE-23 saveTokenz Backlog

Weekly progress digest email → HTML table (Gmail-legible)

VENUS-83 venus Done high

Hub bug: status sender-map hijacks targeted scrp-* DMs. scrp-memsweep-venus matches STATUS_EMITTER_RE (hub.mjs:1815) so a deliberate kind=dm to a specific agent (nw-venus-cc) gets origin_aro stamped aro:hosts-status (1878) + recipients overwritten to the room's 3 nw members (hybrid block 1964-1967). Fix: skip the smap stamp+hybrid overwrite when the send is a deliberate DM to a single explicitly-addressed registered agent (else-branch recipients=[to] at 1937-1940). Keep smap routing for to=all/aro:hosts-status/Elazar-catchall paths (fire-and-forget telemetry). Verify: memsweep DM to nw-venus-cc delivers to ONLY nw-venus-cc; a hostmon-style to=all status still lanes to hosts-status. Bounce coordinated via me (pm).

MSG-254 llmmsg-srv Done high

EFFAUDIT: 2026-07-02 fleet efficiency audit (Elazar brainstorm bs-mr334l5yybu) - dead tunnels/ZT/whey-pointers/keep-alives purge + hub/CDW micro-optimizations

MSG-232 llmmsg-srv Done high

└─ whey+lezama: retire lezama-tun L3 tunnel both sides (whey initiator + lezama /etc/systemd/system/lezama-tun.service; coordinate with nw-lezama-cc) per Elazar decision

MSG-246 llmmsg-srv Done high

└─ whey: delete dead reverse-tunnel units llmmsg-lezama-tunnel.service + llmmsg-venus-hub-localfwd.service (disabled/inactive, superseded by direct HTTPS) + orphaned llmmsg-lezama-zombie-sweep.sh if unreferenced

MSG-233 llmmsg-srv Done high

└─ DECISION(Elazar): retire whey<->lezama L3 tunnel (lezama-tun.service) + whey-vpn-keepalive? Elazar stated nothing whey<->lezama is needed; keepalive also heals GCABA VPN

MSG-244 llmmsg-srv Done high

└─ lezama: tear down dead llmmsg-hub-tunnel.service (ssh -L 9703 to venus; direct HTTPS is live path)

MSG-243 llmmsg-srv Done high

└─ whey: remove dead lezama-watchdog trio (llmmsg-lezama-watchdog svc+timer, supervisor svc+timer, alert svc) + backing scripts lezama-watchdog.sh/lezama-zombie-fix.sh - PM authorizes (llmmsg lane), execution delegated

MSG-234 llmmsg-srv Done high

└─ sh.git: ZT dead-code sweep - llmmsg-doctor.sh zerotier-cli + 172.27.* branch, whey/lezama daemon-probe ZT branches, ccs.sh ZT-fallback hint + stale autossh-watchdog comment

MSG-236 llmmsg-srv Done

└─ whey: retire stale llmmsg-srv.service unit (hub moved to venus) + purge dangling llmmsg-zombie-ca-detector.timer

MSG-235 llmmsg-srv Done

└─ hub.mjs: fix HOST_GUESS map (pm/coder-llmmsgsrv + coder-chatduo now on venus) + drop legacy whey aro_config seed tiers

MSG-238 llmmsg-srv Done

└─ venus: purge zerotier-one package (unit disabled but package installed; ZT retired fleet-wide)

MSG-242 llmmsg-srv Done

└─ venus: retire claude-backup.timer (redundant vs venus-state-backup-daily)

MSG-240 llmmsg-srv Done

└─ CDW: consolidate 4x60s client intervals into one 60s /api/bootstrap tick + strip dead bearerless-whey-hub fallback in server.mjs (coder-llmmsgsrv-cc adds hub-side snapshot endpoint if needed)

MSG-237 llmmsg-srv Done

└─ whey: retire llmmsg-whey-daemon-probe wholesale - service+timer+/etc registry (nw-whey) + remove llmmsg-whey-daemon-probe.sh from sh.git (bin-whey); entire watch-set is dead, timer already inactive - PM approved

MSG-245 llmmsg-srv Done

└─ whey: constrain whey-vpn-keepalive.timer to business hours (Mon-Fri daytime) per Elazar - convenience reconnect only

MSG-247 llmmsg-srv Done

└─ venus: stretch llmmsg-reconcile.timer 10min->30min (safety-net only) - approved by PM

MSG-241 llmmsg-srv Done low

└─ docs: mark zt_* status-probe fields deprecated (keep columns for history)

MSG-239 llmmsg-srv Done low

wi/wi-web: harden against SQLite-direct fallback post-cutover (stub recreation)

BWI-34 basquetwi Done high

404-capture: demote vuln-scanner paths to bot-probe security branch (stop false-paging). GAP: KNOWN_TOP_SEGMENTS fallback keys on first segment only, so any scanner path UNDER a legit top segment (/admin/.git/config, /informes/.env, /practicas/wp-login.php, *.php) inherits own-route -> genuine-bad-route -> alert:true page. No scanner/dotfile denylist exists; X-Audit-Probe demotion only catches our own synthetic probes. FIX: add scanner-path guard BEFORE the classifyOrigin KNOWN_TOP_SEGMENTS fallback (not-found-capture.ts ~L98 pre-filter, or classifyOrigin before not-found-classify.ts:88), demote UNCONDITIONALLY to level=info/category=security + rate-limit like existing bot-probe branch (capture.ts:108-118). Patterns: (a) any path segment matching /^[.]/ EXCEPT .well-known; (b) .php suffix; (c) token denylist wp-admin/wp-login/xmlrpc/phpmyadmin/vendor/cgi-bin/actuator. Triggered by scanner hit /admin/.git/config 2026-06-29 03:38Z id b5625a54.

VENUS-80 venus Done high

/practicas pagination defects (visible at ?state=rectificar): (1) pagination is off/incorrect; (2) the 'show N records' page-size selector is missing. Investigate the list pagination + page-size control, fix both. Elazar 2026-06-30. Reversible UI.

MARS-362 mars Done high

Auto-send welcome email on app-driven access grant/reactivation/login-email-change (no agent intervention)

PLUTO-203 pluto Done high

ccs.sh: persisted LLMMSG_AGENT in shared per-cwd .llmmsg-env overrides .agent-name — cross-agent launch from same cwd poisons resident agent identity

MSG-84 llmmsg-srv Done high

applog-listen hub-eviction recovery: auto re-registration (periodic 5min + 400-triggered immediate). Hub restart evicts listener from roster, re-enqueue loops on 400 until manual bounce. Fix once, redeploy venus/mars/pluto. Proof: hub restart then error delivers real-time w/o manual restart

MSG-82 llmmsg-srv Done high

Whey host-plane DM-leg gaps: clamav EnvironmentFile+ORIGIN_ARO, context-monitor ORIGIN_ARO+script rename, db-size-monitor ORIGIN_ARO

MSG-81 llmmsg-srv Done high

B5 per-student stats page w/ charts (today global-only) — HIGH

VENUS-55 venus Done high

gitpush.sh strict-mode: detect staged deletions in ONLY_PATHS and commit them (git rm --cached silently dropped today)

MSG-109 llmmsg-srv Backlog high

B4 per-student quantitative totals on filter (35 como operador, per-sede) — HIGH client pain

VENUS-54 venus Done high