Control Tower
Engel Operations. The whole board, one screen.
AS OF
09/29/2026, 14:18 CT
source: Problem Ledger + Projects/
09/29/2026, 14:18 CT
source: Problem Ledger + Projects/
959
Open now
3
P0 critical
414
P1 high
542
P2 normal
843
Shipped
16
Projects
Priority queue. P0 and P1
#1464 Kara's hdp-bind-registered-address.ps1 converted Phil's Ethernet from DHCP to Manual 192.168.2.65 with no gateway — ~16h with no internet, recurring
#20260831224738 Chuck (Anti-Gravity) - Two consecutive state hallucinations in conversation
#1742 [chuck-heartbeat-watchdog] 2/2 silence(s) UNFIXED by auto-remediation: engelsplace-fmx-ingest-afternoon, engelsplace-fmx-pm-ingest-afternoon
#1 Tornado safety plan filing (IL HB2987)
#43 PC-to-NAS auto backup rollout — NetBak + Veeam Agent Free, 4 PCs
#54 Power Automate WorkSync Discrete Resend flow — 2 failures past 7 days (Microsoft alert)
#182 NAS philsserver abnormal disk SMART status on bay 1 (3.5" SATA HDD 1) — fired x2 on 6/11
#277 [chuck-drift-guard] 4/5 sections drifted
#301 Live Discord bot token + Resend full-access key + Recall.ai key committed in cleartext across ~11 tracked files (security-sweep lane CHK-LANE-2026-06-27)
#363 [chuck-health-beacon] 1/4 probes failed: version-watch
#366 Live Resend API key committed plaintext in git-tracked IT/status/lane-crossings-log.jsonl:6264 + lane-guard logs raw secrets
#370 [chuck-heartbeat-watchdog] Heartbeat pipeline issue: 1 undelivered alert(s)
#407 [chuck-openai-codex-ecosystem-watch] handler stage failed: spawnSync C:\WINDOWS\system32\cmd.exe ETIMEDOUT
#419 Outcome missing: dreaming-nightly produced no result (verifier could not self-heal)
#430 Schedule Audit Program: verify all ~80 scheduled runs end-to-end + ship deterministic watchdog layer
#441 Desktop Commander shells inherit PATHEXT=.CPL — every external binary (node, pm2, cmd) fails to resolve
#443 chuck-behavior-auditor is BLIND — 5 sessions failed to score (OAuth dead, metered fallback failed)
#456 [chuck-behavior-auditor] handler stage failed: spawnSync C:\WINDOWS\system32\cmd.exe ETIMEDOUT
#483 node-cron MISSES scheduled fires (4 today, incl. the 06:00 engelsplace-youtube-ingest) — missed != failed, so every failure counter and watchdog stays blind
#495 NAS 3 AM auto-REBOOT health gates exist only as English prose in a SKILL.md — no gate code, so missing/unparsed data has undefined behavior
#498 pushAlert marks an alert SENT even when the Telegram POST fails — the outage that breaks the network also swallows the page about it
#504 nas-watch 'silence = healthy' is self-referential — the heartbeat runs INSIDE the detector, so a dead detector is indistinguishable from a healthy NAS
#513 Phil-stated personal facts are not captured to OpenBrain - his full med/supplement protocol was never banked, and Gemini recalled it when we could not
#522 SELF-INFLICTED: wrangler auto-scaffold published Personal_Web_Page/ to a public URL for ~6 min (dead Ghost key exposed)
#528 Painted-pixel count passed while half the promised bars were missing - rendered-output gate was applied superficially one day after being written
#534 auto-remediation.js self-restarts the bot with the BANNED bare pm2 restart (3 sites, live-reachable)
#535 require('./bot.js') boots a second live scheduler - cron registration sits at module scope
#538 Ashley's Opus-tier floor never fires - the Gemini ops path intercepts every interactive reply
#542 /scope hook-enforcement FAILED first live test — Tess bypassed the scope interview at 20:27, 1h47m after Vera's 'scope obedience hook-enforced' receipt on P-00531
#562 MCP stdio server process leak: 421 node procs / 23.9GB RAM on philsgamingmachine
#564 CHUCK_PATCH_TOKEN leaked into an agent transcript by problem.js close failure path
#566 Auto-capture cause-mapper invents a channel/permission accusation for EVERY discord-stage failure — 4th re-audit burned closing its false tickets
#572 Pickup acknowledgments trigger cross-surface watcher ping-pong
#599 Pre-market Robinhood pull reports the PREVIOUS session's move as today's — wrong day-change shipped to Phil in every daily pulse
#632 Jocko intake accepts unauthorized and out-of-order lifecycle transitions
#638 heartbeat-lib.js writes 176KB state non-atomically and silently degrades a torn read to empty
#639 gmail-puller digest failure marks the message seen — meeting never retried, never reaches the 6:15 brief
#667 Ticket-per-fire spam: heartbeat-watchdog + rot-sentinel auto-filers lack open-dup suppression — 2 ongoing incidents wearing 24+ tickets (P-00601..666 cowork-snapshot chain, P-00559/560/561/576/613/655 rot-board chain)
#678 process-snapshot writer archives live secrets verbatim (Cursor crsr_ key in procsnap commandlines)
#688 [fairrite-inbox-watcher] handler stage failed: fetch failed
#696 [fairrite-inbox-watcher] Handler crashed after 8s: fetch failed
#711 Boot-integrity harness covered only 5 of 8 roster agents — Jocko (most important agent), Ashley, Bob untested
#712 Boot-integrity guard reports but does not enforce; its only scheduled executor (chuck-openclaw-on-track-check) has no journal since 2026-07-19
#713 9 boot-tier files over cap — aggregate bootstrap 190880B is 24% over the 150KB budget read every session
#715 Boot injects MISSION-GOSPEL + STANDING-ORDERS fail-open-to-empty — a missing/truncated constitution loads NOTHING with zero signal
#726 Nicole's entire household runs on the GUEST VAP (NETGEAR67/_5G, lanaccess=off) while the MAIN SSID sits unused — structural backdrop to the drop/churn family
#730 engel-ops-bot Discord shard reconnect loop since 09:31 CDT — ws handshake timeouts while outbound HTTPS to discord.com is healthy; suspected blocked event loop in bot process
#733 Telegram inbound: paired-sender allowlist is written but never enforced — any sender routes into askAgent
#745 Crosstalk builder-response watcher latched silent 9 days — one-shot alert flag muted 30+ Builder replies (FIXED+falsified this pass)
#746 MISTAKE-PROOFING PROGRAM (Phil-authorized 2026-07-25): move up the poka-yoke hierarchy across all 8 surfaces — stop adding detectors
#760 PM2Heartbeat auto-recovery runs the BANNED bare 'pm2 restart engel-ops-bot' unattended every 30 min — the exact command P-00507 forbids, in our own self-heal path
#772 [chuck-registration-guard] Handler returned critical result (overall=red): 🚨 Scheduled-task registration guard: UNDETERMINED
Could not read the inputs, so registration integrity is UNKNOWN — this is not a pass.
• no "[cron] N task(s)
#788 Work-registration gate enforces NOTHING in production — 86/87 intake registrations untagged; any agent's registration in the 4h weak window satisfies any session
#789 Two standing laws (escalate-through-Jocko, Battle-Buddy-grade) had NO mechanical gate — a per-response formatting habit overrode the constitution
#790 Chain of command: ask-block doctrine still addressed asks to Phil in two Tier-1 files after ICAR-05 claimed the fix
#793 Battle-buddy + registration Stop gates are BLIND to every file outside C:/Users/engelp/Claude_Lives_Here — including the entire live engelsplace working copy
#794 jocko-intake.js hard-limits graders to jocko|vera — no troop can record a Battle Buddy grade, and Jocko has 6 self-graded A's in the live ledger
#799 jocko-intake.js has no reopen verb — a bad GRADE cannot be corrected, only a bad close can
#802 battle-buddy-gate hands you a remediation command that would falsely grade a DIFFERENT agent's open work item
#806 Signal-with-no-consumer CLASS: digest staging (fixed 07-26), lane-crossings 4MB, preflight-gate-log, commit-pressure-watch, run-hidden-errors, behavior-auditor output dead 3d
#807 run-hidden.vbs swallows child exit codes — LastTaskResult=0 unconditionally for ~15 Windows tasks (false-green class)
#809 CA-6 STALE-TICKET ESCALATION — the only corrective action aimed at the behavioral root has existed for 2 days as a red row in a document with no ledger ticket
#810 gmail-puller bypasses the build gate — one bad frontmatter file pushes straight to main and freezes the ENTIRE site, every collection, with nothing watching
#813 pm2 reconciler: LIVE_NOT_DECLARED — avi-csn-watcher
#814 crosstalk-watcher had no structural separation between test and production invocation - a unit test spawned a live authenticated Jocko session against a temp fixture
#823 problem.js silently swallows unknown flags: a mistyped --severity downgrades a ticket to P2 and BYPASSES the mandatory ICAR-at-discovery gate
#825 Work identity is resolved from the SESSION, not the WORK — the gate binds unrelated edits to a foreign WI and inherits its redlined state (re-cut of P-00819 at the intended P1)
#826 Grading is FIRST-GRADER-WINS and terminal — a second grader's dissent is structurally unrecordable, so the finding that matters most is the one most likely to be lost
#840 CSN watchers silently drop ALL SURFACES broadcasts that carry a co-addressee before the arrow
#842 avi-csn-watcher has NEVER delivered anything — it scans bottom-up for the newest entry but CSN is prepend-newest
#852 Three Cloudflare-side writers commit straight to engelsplace main via the GitHub Contents API with NO validation — the half of P-00810 that no local build gate can reach
#863 14 of 30 due bot-cron tasks silently never fired — every task at minute :00 or :30 is dead across 6 lanes, incl. the backup verifier that owed an alarm on a REAL outage
#865 [kara-hdp-backup-verifier] Handler returned critical result (overall=critical): HDP Backup Verifier: CRITICAL
1 backup problem(s) across 2 host(s), 1 healthy: philsgamingmachine — job a863c924-31d1d623-70b83a40d896 STALE: last success 2026-
#869 ICAR overdue CA: ICAR-2026-06-16-02#CA-1 — ICAR-2026-06-16-02 CA-1
#870 ICAR overdue CA: ICAR-2026-06-16-02#CA-2 — ICAR-2026-06-16-02 CA-2
#871 ICAR overdue CA: ICAR-2026-06-16-02#CA-3 — ICAR-2026-06-16-02 CA-3
#872 ICAR overdue CA: ICAR-2026-06-19-01#CA-1 — ICAR-2026-06-19-01 CA-1
#873 ICAR overdue CA: ICAR-2026-06-19-01#CA-2 — ICAR-2026-06-19-01 CA-2
#874 ICAR overdue CA: ICAR-2026-06-20-02#CA-1 — ICAR-2026-06-20-02 CA-1
#875 ICAR overdue CA: ICAR-2026-06-20-02#CA-2 — ICAR-2026-06-20-02 CA-2
#877 ICAR overdue CA: ICAR-2026-06-20-02#CA-4 — ICAR-2026-06-20-02 CA-4
#880 ICAR overdue CA: ICAR-2026-06-20-03#CA-2 — ICAR-2026-06-20-03 CA-2
#881 ICAR overdue CA: ICAR-2026-06-20-03#CA-3 — ICAR-2026-06-20-03 CA-3
#882 ICAR overdue CA: ICAR-2026-06-22-02#CA-5 — ICAR-2026-06-22-02 CA-5
#883 ICAR overdue CA: ICAR-2026-06-24-02#CA-3 — ICAR-2026-06-24-02 CA-3
#884 ICAR overdue CA: ICAR-2026-06-24-02#CA-4 — ICAR-2026-06-24-02 CA-4
#885 ICAR overdue CA: ICAR-2026-06-27-03#CA-2 — ICAR-2026-06-27-03 CA-2
#886 ICAR overdue CA: ICAR-2026-06-27-04#CA-2 — ICAR-2026-06-27-04 CA-2
#887 ICAR overdue CA: ICAR-2026-06-27-04#CA-3 — ICAR-2026-06-27-04 CA-3
#888 ICAR overdue CA: ICAR-2026-06-28-01#CA-2 — ICAR-2026-06-28-01 CA-2
#889 ICAR overdue CA: ICAR-2026-06-28-01#CA-3 — ICAR-2026-06-28-01 CA-3
#890 ICAR overdue CA: ICAR-2026-06-28-01#CA-4 — ICAR-2026-06-28-01 CA-4
#891 ICAR overdue CA: ICAR-2026-07-05-01#CA-4 — ICAR-2026-07-05-01 CA-4
#894 ICAR overdue CA: ICAR-2026-07-25-02#CA-8 — ICAR-2026-07-25-02 CA-8
#897 icar-ca-consumer reads 32 of 103 ICARs and reports the 71 it cannot read as a clean scan — no denominator, empty dir exits 0/OK
#905 No cross-surface watcher keeps a PER-ENTRY body fingerprint, so an amended field order is undeliverable — two watchers hold only a single scalar fingerprint
#907 Ashley's release gate is unfalsifiable: jocko-intake.js CARVED_OUT still refuses her, so no grade she ships outbound on is ever recorded
#908 ICAR overdue CA: ICAR-2026-07-21-03#CA-3 — ICAR-2026-07-21-03 CA-3
#909 ICAR overdue CA: ICAR-2026-07-21-03#CA-4 — ICAR-2026-07-21-03 CA-4
#914 crosstalk-watcher PICKUP ACK works but authenticated claude-jocko EXECUTION_START exits 1 (closeout-gate + spend limit)
#916 [kara-tunnel-reachability] handler stage failed: tunnel check produced no fresh artifact (mtime 2026-07-28T07:00:01.711Z predates run start) — child likely died pre-writ
#925 ICAR overdue CA: ICAR-2026-07-20-01#CA-2 — ICAR-2026-07-20-01 CA-2
#926 ICAR overdue CA: ICAR-2026-07-20-01#CA-3 — ICAR-2026-07-20-01 CA-3
#927 ICAR overdue CA: ICAR-2026-07-25-02#CA-5 — ICAR-2026-07-25-02 CA-5
#928 ICAR overdue CA: ICAR-2026-07-26-01#CA-1 — ICAR-2026-07-26-01 CA-1
#933 pm2 reconciler: DECLARED_NOT_LIVE — avi-csn-watcher
#934 pm2 reconciler: DECLARED_NOT_LIVE — bot-crashloop-guard
#935 pm2 reconciler: DECLARED_NOT_LIVE — config-guardian
#936 pm2 reconciler: DECLARED_NOT_LIVE — crosstalk-watcher
#937 pm2 reconciler: DECLARED_NOT_LIVE — engelsplace-dev
#938 pm2 reconciler: DECLARED_NOT_LIVE — icar-ca-consumer
#939 pm2 reconciler: DECLARED_NOT_LIVE — openbrain-transcript-watcher
#940 pm2 reconciler: DECLARED_NOT_LIVE — tard-notes-watcher
#941 pm2 reconciler: DECLARED_NOT_LIVE — vera-notes-watcher
#946 problem.js silently TRUNCATED multi-line ticket notes on every update — 104,299 chars destroyed across 35 tickets
#957 Outcome missing: kids-daily-drop produced no result (verifier could not self-heal)
#964 Vera's Codex boot bundle silently truncates: 5 of 16 files never arrive, including every system-wide constraint file (CROSS-SURFACE-NOTES, ORG_STATE, SYSTEM_STATE, memory/AGENTS.md, decisions-log)
#968 marek-quote-watcher Gmail sweep gates on sender domain, silently dropped TWO Evexia replies (07-26 and 07-29) and produced a false 'Evexia has never replied' record
#971 SELF-INFLICTED: I leaked the live CHUCK_PATCH_TOKEN into the ops ledger by pasting an unescaped command-substitution into a problem.js note
#973 Pre-open snapshot's prev_close_equity holds the session BEFORE the last close — P-00599's shifted INPUT survived the CA-1 fix and is still quotable
#975 18 live scripts in IT/scripts are UNTRACKED — including the per-turn prompt hook and the zombie self-term generator — and load-bearing-paths.json does not name them
#978 crosstalk-watcher spawns a CHUCK responder for JOCKO-addressed Builder blocks: today's P-00969 fix widened the trigger and left the launcher+prompt on the retired seat — 3 OPEN blocks orphaned tonight, one owner-directed
#980 ICAR overdue CA: ICAR-2026-07-21-03#CA-5 — ICAR-2026-07-21-03 CA-5
#981 ICAR overdue CA: ICAR-2026-07-21-03#CA-6 — ICAR-2026-07-21-03 CA-6
#982 ICAR overdue CA: ICAR-2026-07-25-02#CA-3 — ICAR-2026-07-25-02 CA-3
#983 ICAR overdue CA: ICAR-2026-07-25-02#CA-6 — ICAR-2026-07-25-02 CA-6
#984 ICAR overdue CA: ICAR-2026-07-25-02#CA-11 — ICAR-2026-07-25-02 CA-11
#985 ICAR overdue CA: ICAR-2026-07-26-05#CA-13 — ICAR-2026-07-26-05 CA-13
#986 ICAR overdue CA: ICAR-2026-07-26-06#CA-1 — ICAR-2026-07-26-06 CA-1
#987 ICAR overdue CA: ICAR-2026-07-26-06#CA-2 — ICAR-2026-07-26-06 CA-2
#988 ICAR overdue CA: ICAR-2026-07-27-04#CA-1 — ICAR-2026-07-27-04 CA-1
#989 ICAR overdue CA: ICAR-2026-07-27-04#CA-2 — ICAR-2026-07-27-04 CA-2
#990 CROSSTALK.md cross-host writes are last-writer-wins with NO lock and NO conflict copy — my 18:34:47 write silently destroyed a 4KB Builder block posted at 18:34:34
#993 Jocko graded a daemon fix A on source-side evidence while the running pm2 image still held the retired code for 5h19m, and the routed cycle decision that would have caught it sat unanswered inside the ticket he graded from
#994 Jocko reproduced a subordinate's truncated quotation verbatim into the permanent external CROSSTALK record and built an affirmative concession on it, inside the block whose sole purpose was correcting a misgrade
#998 No agent launcher injects <AGENT>_PATCH_TOKEN, so every agent must grep a credentials file on the command line to file a ticket — the exact transcript-exposure cost P-00898 was designed to eliminate, and the reason real findings go unticketed
#1010 jocko-intake grade accepts any --grader value from any caller: an agent can record an A in another agent's name, and I just did it to my own work
#1017 battle-buddy-gate weak time-window matcher named the WRONG agent's work item, and a live grade was recorded against it: Vera's WI-20260730023936814dq7-vera is terminally COMPLETE with an A whose notes describe Ashley's work
#1018 CA-6 refuses every grade recorded by a deployed *-worker sub-agent, because a sub-agent inherits its parent CLAUDE_CODE_SESSION_ID - the Battle Buddy grade path is closed for the third time
#1036 Two Scheduled SKILL.md trees, no guard - an ICAR corrective was applied to the copy nothing reads and both of Chuck daily runs booted without the gates file
#1038 CROSS-SURFACE-NOTES.md is 1,007,559 bytes (984 KB) — a declared every-boot read that is 49x the 20KB bootstrap cap, and the likely cause of P-00964's silent Codex truncation
#1039 verify-marketplace-clean: 13 of ~16 user-scope plugin installs have NO .install-manifests entry (kara, alex, jocko, ashley, bob, scope, battle-buddy, loop-closer, systems-check, build-plugin, bridge-sync, cursor-architect, animated-scroll-site) + 1 dead marketplace entry
#1043 kara-network-watch graded a POWERED-OFF NAS as reachable for 66h — ping parser trusted the '(0% loss)' summary line, which Windows prints identically for a dead host
#1047 No Cloudflare Workers deploy credential AND no CI path — two engelsplace Workers run ungated code with manual wrangler as the only route
#1050 credentials-ledger.md omits the Cloudflare WORKERS deploy token — the ledger is the 'grep before any access ask' authority, and a gap in it made me route a false blocker
#1054 80 scheduled-task sessions across 19 tasks were killed at turn 1 by a spend-limit refusal 07-25..07-29, and lastRunAt was stamped for every one of them
#1068 [dawn-reply-watch] responder ran but never confirmed — possible partial send
#1069 [dawn-reply-watch] Dawn wrote; responder unconfirmed (P-01068) — held for review
#1076 Dawn responder cannot SEND unattended - permission classifier blocks the send script on every shell
#1096 [dawn-reply-watch] Handler crashed after 15s: fetch failed
#1098 [dawn-reply-watch] handler stage failed: fetch failed
#1100 ICAR overdue CA: ICAR-2026-07-21-03#CA-8 — ICAR-2026-07-21-03 CA-8
#1101 ICAR overdue CA: ICAR-2026-07-21-03#CA-9 — ICAR-2026-07-21-03 CA-9
#1102 ICAR overdue CA: ICAR-2026-07-25-02#CA-12 — ICAR-2026-07-25-02 CA-12
#1103 ICAR overdue CA: ICAR-2026-07-26-01#CA-6 — ICAR-2026-07-26-01 CA-6
#1104 ICAR overdue CA: ICAR-2026-07-26-04#CA-2 — ICAR-2026-07-26-04 CA-2
#1105 ICAR overdue CA: ICAR-2026-07-26-05#CA-9 — ICAR-2026-07-26-05 CA-9
#1106 ICAR overdue CA: ICAR-2026-07-26-05#CA-14 — ICAR-2026-07-26-05 CA-14
#1107 ICAR overdue CA: ICAR-2026-07-27-04#CA-4 — ICAR-2026-07-27-04 CA-4
#1110 avi-debt-profiler bot-cron stale — last ran ~327h ago vs weekly Sunday cadence
#1115 NETGEAR MS108EUP admin password printed into an agent transcript — a redaction filter was written against a label the credential file does not use, and was never tested against the real file
#1124 [chuck-schedule-snapshot] Failed to write snapshot: UNKNOWN: unknown error, open 'C:\Users\engelp\Claude_Lives_Here\IT\status\schedule-snapshot.json'
#1127 system-health-monitor is hosted NOWHERE since 2026-07-19 and SYSTEM_STATE listed it ACTIVE — the daily 5:15 AM health roll-up silently stopped existing
#1130 Peptide dose calculator gives THREE impossible fixes when the dose exceeds the whole vial — live on ss-31, tirzepatide and retatrutide
#1138 Bot scheduler drops ~45 cron fires/8h — execSync pm2 jlist in auto-remediation blocks the event loop; the watchdog manufactures the silence it reports
#1139 Outcome missing: chuck-schedule-snapshot produced no result (verifier could not self-heal)
#1143 battle-buddy-gate names another agent's work item on a self-declared weak match, and prescribes the grade command that would forge a record onto it
#1145 engel-ops-bot STOPPED cleanly (pid 0, restarts 0) between 07:48 and 08:09 CDT 2026-07-31 — no crash signature, no cause in any log; every bot-cron monitor is dark while it is down
#1146 pm2 reconciler: PROCESS_DOWN — engel-ops-bot
#1150 A 3-day account spend-limit blackout killed ~152 sessions across 19 scheduled-task families (07-27..07-29) and NO control detected it as one event — including 2 silent nights for the kids' channels
#1151 Every PowerShell Discord bot-API alert has failed silently since 2026-04-24: no User-Agent header returns 403 code 40333, whose text 'internal network error' reads as a transient blip
#1154 Daily digest cannot retract a resolved alert or dedup near-identical Phil-asks: a self-healed condition staged red reaches Phil as LIVE up to 24h later, and two near-duplicate asks both shipped in one brief
#1156 Our spend telemetry watches the WRONG METER — it reported $0.19 and 'ok' straight through the 3-day blackout and is still reporting healthy today
#1159 I truncated Network/WHERE-WE-ARE.md to ZERO BYTES with open(path,'w') — a mid-write exception destroyed 150KB of boot-critical continuity; recovered from git
#1163 ICAR escalation ladder is doctrine with no executor: 10 open ICARs are 7+ days past target and owe a mandatory URGENT push nobody sent, and 14 open ICARs carry no target date at all so they can never be flagged
#1164 chuck-problem-auto-closer's heartbeat is 7.25 days stale and RED while its own output artifact is 5 hours fresh — the liveness signal and the work disagree by a week
#1166 Action items never expire and nothing re-verifies their premise — live-open-actions.js --verify asserts a freshness it does not establish, and fed Phil a 3-month-dead ask
#1167 [CA-3] Dawn responder: the 'send permission cleared' event has no code consumer
#1171 pm2-heartbeat-enhanced.ps1:270 issues the bare 'pm2 restart engel-ops-bot' that P-00507 bans - unattended, every 30 min, and it is what recreated the process at 2026-07-31T13:18:52Z with no receipt
#1173 bot-crashloop-recover.js writes its receipt only AFTER pm2 start - a run killed between step 1 (pm2 stop) and step 4 leaves the bot STOPPED with zero record it ever ran; caused a 16m11s unreceipted bot outage on 2026-07-31
#1174 Grading backlog root cause: a grade the integrity guard REFUSES is routed to a NAME, not delivered to a QUEUE - the verdict is silently lost and the work reads as never-graded forever
#1175 loop-closer over-merged 47 distinct ICAR corrective actions into one recurrence class — the dedup fix's own dedup was the next defect
#1176 Battle Buddy gate is silently banking ungraded work: its give-up path writes to ungraded-work.jsonl, which has no consumer
#1182 Telegram bombarded Phil with commands and P-TEST traffic - no payload gate on his phone
#1183 Chart correction fixed the file but left the superseded dose live in the OpenBrain retrieval layer
#1187 ICAR overdue CA: ICAR-2026-07-21-03#CA-7 — ICAR-2026-07-21-03 CA-7
#1188 ICAR overdue CA: ICAR-2026-07-26-01#CA-2 — ICAR-2026-07-26-01 CA-2
#1189 ICAR overdue CA: ICAR-2026-07-26-01#CA-3 — ICAR-2026-07-26-01 CA-3
#1190 ICAR overdue CA: ICAR-2026-07-26-06#CA-3 — ICAR-2026-07-26-06 CA-3
#1191 ICAR overdue CA: ICAR-2026-07-27-04#CA-3 — ICAR-2026-07-27-04 CA-3
#1192 ICAR overdue CA: ICAR-2026-07-27-06#CA-3 — ICAR-2026-07-27-06 CA-3
#1193 ICAR overdue CA: ICAR-2026-07-27-06#CA-5 — ICAR-2026-07-27-06 CA-5
#1194 ICAR overdue CA: ICAR-2026-07-29-07#CA-4 — ICAR-2026-07-29-07 CA-4
#1195 ICAR overdue CA: ICAR-2026-07-29-08#CA-1 — ICAR-2026-07-29-08 CA-1
#1196 ICAR overdue CA: ICAR-2026-07-29-08#CA-2 — ICAR-2026-07-29-08 CA-2
#1198 pushAlert() is a WILDCARD around the registered-sender gate: any caller with no source is auto-authorized at emergency tier
#1199 verify-telegram-senders.js CANNOT FAIL on the pushAlert lane - it collects pushAlert sources for hygiene but never validates one
#1200 The 2026-07-31 payload gate has NO emergency fail-safe - a genuine emergency naming a Windows path dies into the digest
#1204 The 6h duplicate suppressor BREAKS the Telegram regression gate: a second run inside 6 hours reports 15 false failures, including a false 'Phil could go dark'
#1207 ashley-health-puller normalizer filed after-midnight data on the wrong day and overwrote multi-session sleep (exit 0, silently wrong)
#1211 A half-finished publish leaves ONE unstaged registry file and silently jams EVERY engelsplace pipeline push - live site went 24h and 310 commits stale with nothing obligated to act
#1215 AGENT_BOARD ask rows assert a live condition but nothing re-tests them, so an out-of-band resolution ages into a false present-tense claim to Phil
#1217 Tard and Avi CSN watchers deliver grade requests to inbox but never execute — grades stall until a human opens that surface
#1219 ICAR overdue CA: ICAR-2026-07-26-01#CA-7 — ICAR-2026-07-26-01 CA-7
#1220 ICAR overdue CA: ICAR-2026-07-27-04#CA-5 — ICAR-2026-07-27-04 CA-5
#1221 ICAR overdue CA: ICAR-2026-07-27-04#CA-6 — ICAR-2026-07-27-04 CA-6
#1222 ICAR overdue CA: ICAR-2026-07-29-06#CA-1 — ICAR-2026-07-29-06 CA-1
#1223 ICAR overdue CA: ICAR-2026-07-29-08#CA-3 — ICAR-2026-07-29-08 CA-3
#1224 ICAR escalation: ICAR-2026-06-16-02 corrective-action target missed by 39d
#1225 ICAR escalation: ICAR-2026-06-20-03 corrective-action target missed by 35d
#1226 ICAR escalation: ICAR-2026-06-20-04 corrective-action target missed by 35d
#1227 ICAR escalation: ICAR-2026-06-22-02 corrective-action target missed by 33d
#1228 ICAR escalation: ICAR-2026-06-24-01 corrective-action target missed by 31d
#1229 ICAR escalation: ICAR-2026-06-24-02 corrective-action target missed by 31d
#1230 ICAR escalation: ICAR-2026-06-27-04 corrective-action target missed by 28d
#1231 ICAR escalation: ICAR-2026-06-28-01 corrective-action target missed by 27d
#1232 ICAR escalation: ICAR-2026-07-05-01 corrective-action target missed by 27d
#1233 63 of 123 ICARs declare no Status at all, so 'open ICARs' is not an enumerable population and every ICAR-level control runs only on the half of the library that happens to be well-formed
#1237 verify-page-render.js silently dropped every repeated --expect-text/--expect flag and still printed RESULT: PASS
#1238 battle-buddy-gate.js reports 'no grade recorded' for work that WAS graded, because it cannot see verdicts jocko-intake deliberately queued
#1241 The semiconductor-paper citation I fixed on ss-31 on 07-30 was still live on epithalon two days later, printing a 31 percent mortality figure that exists in no source
#1242 I published two ungraded pages with a wildcard git add, in the same session I spent defending that publish gate
#1243 LIVE Chuck Telegram bot token hardcoded in TWO TRACKED throwaway test scripts - byte-identical to the production notifier credential
#1244 OpenBrain backup verification chain is keyed to a RETIRED artifact class - 42 days of SQLite backups have zero restore proof and P-00156 is a false-label resolved
#1248 Tard Cursor wake arm silently died for 1+ day while state file still said armed=true
#1250 Pages half-corrected: a claim gets fixed in one block and left standing in three others, and nothing on the site can detect it
#1251 Verdict-queue file's Verdict line never updates on a later grade attempt — a superseded B masquerades as current while the real A hides in raw JSONL
#1252 Playwright browser binary is missing — verify-page-render.js cannot launch, so the mandated render gate is currently unrunnable
#1253 An erroneous redline cannot be vacated — retract requires a prior graded event, so a false accusation strands the WI and only the wrongly-accused worker can unjam it
#1255 A same-session grade refusal enqueues a cross-grade obligation that nothing enforces — the worker published on the refused self-grade while the obligation stayed unsigned
#1256 ICAR id collisions: daily gate registered INFO 'not alarmed', 7 reports share 3 ids on 2026-08-01
#1258 Wrong-root memory cite class fixed at 2 instances and never bounded: 8 live operative sites remain outside the memory tier and no gate detects the class
#1259 CSN wipe recovery reconstructed 4 of 6 lost entries with no denominator declared; the 2 dropped included a release-gated grade order
#1266 [insurance-reply-watch] Handler crashed after 75s: fetch failed
#1267 [insurance-reply-watch] handler stage failed: fetch failed
#1270 ALL FIVE watcher scripts export nothing and have no require.main guard, so require()-ing one to reach its receipt gate boots a SECOND live watcher that spawns authenticated sessions
#1272 ICAR ID collision recurrence: 3 IDs / 7 documents survived ~22h — all hand-filed, bypassing the collision-proof allocator
#1273 insurance-reply-watch pages 13x overnight for a candidate it classed as noise, and names a script (IT/scripts/insurance.js) that does not exist
#1275 Bootstrap/agents/memory-tier janitor findings evaporate: report written to a GITIGNORED file that every run overwrites, and tickets_filed=0 — the code-sweep tier of the same program files tickets
#1276 Competing boot protocols: ORG_STATE.md and .agents/AGENTS.md order agents to boot without ever naming MISSION or OpenBrain — memory/README.md was fixed at the instance yesterday and the class was never swept
#1278 nas-watch's HDP backup check reported 'all green' on a 7-day backup outage — zero completed backups was an explicit green path, and the blind spot went load-bearing when the co-occurring VSS-cap alarm cleared
#1280 pm2-staleness-checker reports 'All expected PM2 processes are online and healthy' and exits 0 when 'pm2 jlist' returns EMPTY stdout — zero services verified, verdict byte-identical to a healthy fleet
#1284 system-health-monitor orphaned by the JOC-0024 migration - MCP-auth and LAN-ping coverage dark 14 days
#1289 daily-financial-report has NEVER fired since the 07-26 migration — 6 due weekdays, 0 fires — and the heartbeat watchdog cannot see it because it derives its watch list from execution records
#1293 jocko-grade-timeout-failover posts 'DIAGNOSE WHY YOU DID NOT PULL' entries addressed to JOCKO with no executor: 28 in CROSS-SURFACE-NOTES, zero ever answered
#1298 [openbrain-watchdog] handler stage failed: Command failed: node IT/scripts/openbrain-watchdog.js
node:internal/modules/cjs/loader:1459
throw err;
^
Error:
#1302 OpenBrain autocapture banks Anti-Gravity sessions as EMPTY shells — _extract_text only understands the Claude Code schema
#1308 Auto-disable is permanent-until-manual, and the operator alert instructs the ONE recovery action that provably does not work
#1316 Two enabled scheduled tasks have NEVER executed since the 2026-07-26 Cowork->bot-cron migration: daily-financial-report + john-weekly-compliance-update
#1322 Grade recording is forgeable at the root: the drainer records under a constant session id, neutralizing intake's only unforgeable guard
#1323 Grade-provenance guard degrades to a pass when the worker registers with no session id (sibling of P-01322)
#1325 jocko-intake done-gate accepts a 'done' with no readable artifact — a turn-only deliverable becomes gradeable and lockable
#1326 pm2 reconciler: LIVE_NOT_DECLARED — grade-dispatch-watcher
#1328 Fast-grade dispatcher JAMS at its 2-slot concurrency cap when a spawned grader hangs and never releases its slot — all new grades wait forever (silent waiting Phil banned)
#1329 Cursor/Tard grades record actorCheck=no-recording-session, and claude-tard.cmd mints its own session id — non-Code grades have NO provenance
#1333 One hand-written content file jams EVERY engelsplace push, and nothing is obligated to fix it — 94 commits / 18.5h stale site
#1334 Battle Buddy v8.0 provenance token forgery — two forged verdicts written into the production intake ledger (re-filed from a clobbered slot)
#1335 Gemini spend cap BREACHED - 17.50 vs 15.00 cap, API blocked; ~17.49 from an UNMETERED consumer
#1339 Nobody can tell who overwrote a ticket: five surfaces write to one content tree and only ONE of them records an actor
#1343 [heartbeat-delivery] #it-ops heartbeat alerts are not reaching Discord (3 consecutive fails)
#1344 [kara-tunnel-reachability] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1345 ICAR overdue CA: ICAR-2026-06-20-02#CA-8 — ICAR-2026-06-20-02 CA-8
#1346 ICAR overdue CA: ICAR-2026-07-26-01#CA-8 — ICAR-2026-07-26-01 CA-8
#1347 ICAR overdue CA: ICAR-2026-07-26-04#CA-3 — ICAR-2026-07-26-04 CA-3
#1348 ICAR overdue CA: ICAR-2026-07-29-01#CA-4 — ICAR-2026-07-29-01 CA-4
#1349 ICAR overdue CA: ICAR-2026-07-29-05#CA-4 — ICAR-2026-07-29-05 CA-4
#1350 ICAR overdue CA: ICAR-2026-07-29-05#CA-6 — ICAR-2026-07-29-05 CA-6
#1351 ICAR overdue CA: ICAR-2026-07-29-07#CA-5 — ICAR-2026-07-29-07 CA-5
#1352 ICAR overdue CA: ICAR-2026-07-29-07#CA-6 — ICAR-2026-07-29-07 CA-6
#1353 ICAR overdue CA: ICAR-2026-07-29-08#CA-4 — ICAR-2026-07-29-08 CA-4
#1354 ICAR overdue CA: ICAR-2026-08-02-01#CA-4 — ICAR-2026-08-02-01 CA-4
#1355 ICAR escalation: ICAR-2026-07-20-01 corrective-action target missed by 7d
#1356 ICAR escalation: ICAR-2026-07-20-02 corrective-action target missed by 7d
#1357 [kara-hdp-backup-verifier] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1358 [engelsplace-fmx-ingest-midday] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1359 [engelsplace-fmx-pm-ingest-midday] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1360 [engelsplace-youtube-ingest] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1361 [engelsplace-fmx-ingest-afternoon] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1362 [engelsplace-fmx-pm-ingest-afternoon] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1364 [chuck-daily-ops-report] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1365 Grade-dispatch SLA rescan spin burned Cursor Pro+ (52 spawns per WI; no-grader firehose)
#1366 [alex-trading-desk-pilot] Handler returned critical result (overall=error): 📊 Trading-desk pilot run FAILED — Command failed: python -X utf8 C:\Users\engelp\Claude_Lives_Here\Finance\trading-desk-pilot\run_pilot.py
(log-only pilot; n
#1371 jocko-intake same-session grade guard fires on a SESSION-ID COLLISION and refuses an independent grader
#1372 2026-08-04 mass deletion: repo Scheduled/ SKILL.md (8 files) + entire out-of-repo ~/.claude/scheduled-tasks/ tree destroyed while live bot crons depend on both — incl. Dawn responder contract RESPOND.md
#1373 Intake queue state machine strands work: 148 items stuck at 'started' (78 with no request-context), auto-registered items never reach 'done', graders mechanically blocked
#1377 Gemini spend-cap burn: Avi automation ran the CLI's PAID default model unpinned for 2 days — the Gemini twin of the 136-spawn Cursor burn
#1378 Lean Cleanup 2026-08-04 (Jocko's own) disabled 28 bot crons + 29 native routines including every watchdog that would have detected it — Phil's daily brief dark since 08-02, nightly git safety snapshot off
#1382 [chuck-registration-guard] Handler returned critical result (overall=red): ⚠️ Scheduled-task registration drift (live boot 2026-08-06T02:55:50, 52 registered)
• DEREGISTERED jocko-grade-timeout-failover — succeeded 3.7d ago but is NOT
#1384 [chuck-rot-sentinel] nightly: 5/6 checks failing after auto-heal (board, boot-caps, plugins, hooks, watchers)
#1385 [openbrain-watchdog] handler stage failed: unknown handler "openbrain-watchdog" — add it to SCHEDULED_HANDLERS in bot.js
#1391 [chuck-heartbeat-watchdog] 1 auto-disabled task(s) need Chuck
#1392 Process memory leak: asus_framework
#1399 ICAR-2026-07-30-04 CA-3's deterministic consumer has been switched off since 2026-08-04 — and the first finance failure after it went dark was never reported
#1401 conversation-index produced 0 chunks for 5 days while reporting success (better-sqlite3 MODULE_NOT_FOUND)
#1402 Phil's daily brief replayed a 4-day-old schedule-watchdog snapshot as today's outages
#1404 11 monitored surfaces went stale in the 2026-08-03/04 cleanup window and nothing is watching, because the watchdog itself is disabled
#1412 [chuck-heartbeat-watchdog] 4 undelivered alert(s)
#1424 Claude Code routine registry is MEMORY-authoritative: an on-disk enabled=true edit is inert and gets silently clobbered by the running Desktop app
#1425 ashley-health-pull-daily silently did not run for ~6 days — Aug 4-7 absent from the chart, Aug 1-3 frozen at partial values, nothing detected it
#1428 John was transferred off Engel Ops 2026-08-04 and nothing reassigned his obligations: 21 live references (9 open tickets incl 2 P1, an enabled Sunday bot cron, a Claude Code daily task, the weekly audit rotation, the roster) still point at a persona whose files are deleted
#1431 Crypto loop's bare 'except Exception: continue' published crypto_value $0.00 with BTC still held, exit 0 — a log-encoding bug presenting as a missing position (P-00747 recurrence)
#1432 Battle Buddy v4.2 logic-pass doctrine change never swept into jocko-verdict-queue-drain recorder — jammed all grade=A recording
#1447 [chuck-heartbeat-watchdog] 12 undelivered alert(s) · 1 auto-disabled task(s) need Chuck
#1449 [chuck-registration-guard] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1451 [control-tower-daily-brief] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1453 [sb-gratitude-prompt] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1456 [chuck-heartbeat-watchdog] 8 undelivered alert(s) · 1 retired zombie(s) auto-fixed + 3 auto-disabled task(s) need Chuck
#1457 [chuck-openai-codex-ecosystem-watch] discord stage failed: getaddrinfo ENOTFOUND discord.com
#1458 [sb-gratitude-checkin-watch] Handler returned critical result (overall=error): SB gratitude check-in watch: could not read #champions-chat — getaddrinfo ENOTFOUND discord.com
#1460 Claude Code scheduled-task sessions ignore the settings.json model pin and boot on Opus 5, against Phil's 2026-08-06 RED order
#1461 The weekly internal deep-dive audit task was deleted on 2026-08-04 with no successor on any surface, and the 08-10 restore was signed off by a control that cannot see a removed capability
#1463 Nicole's router (192.168.2.2) DHCP lease was 1440 SECONDS (24 min) — Phil's PC lost its address on every renew
#1466 Tard Cursor boot files bury CROSS-SURFACE-NOTES and persona boot skips it
#1469 Scheduled tasks auto-disabled by post-work Discord-delivery DNS blip — primary handler had succeeded
#1473 gemini-cost-rollup.js estimate-only run claimed a real-billed source and wiped the last real MTD
#1474 Revoking kara_gaming_admin_ed25519 disarmed 9 elevated-loopback consumers unswept — the P-00509 VSS cap remediation AND its detector are both dead
#1477 Both FMX afternoon pulls missed their 14:45 fire on single-value crons, auto-remediation failed, and NO ledger ticket was created — a missed run is invisible where a failed run is not
#1479 Fair-Rite's own weekly operating record is parsed, published to a website, and consumed by NO agent — the compliance officer built a program from assumptions the record refutes
#1480 Unattended Cursor CLI spawn storms had no halt — only a launcher rate cap
#1483 ashley-health-puller exits 1 (=dead credential) on a transient file-write lock — a healthy feed pages Phil to re-mint a token that just refreshed
#1486 Grade backlog structurally accumulates: 84 WIs, 79 done-never-graded, 43 auto-registered, over 2 weeks
#1492 [chuck-app-scheduler-twin-guard] Handler returned critical result (overall=fail): ⚠️ App-scheduler twin guard FAILED (exit 0)
{
"ok": true,
"apply": true,
"registry": "C:\\Users\\engelp\\AppData\\Roaming\\Claude\\claude-code-sessions\\
#1493 Scheduled-task Stop hook false-trips every successful Claude cron and spawns Cursor
#1498 Jocko has no forcing function to PROCESS submitted missions — 8 rotted overdue while their deliverables sat graded-A
#1500 Catch-up net only recovered daily crons so interval watchdogs stayed dark after a missed tick
#1501 plugin-build builder has no plugin.json gate — ships the exact 'missing plugin.json' install error the pipeline exists to prevent
#1502 plugin install-verifiers hardcode their roster — 10 of 16 installed plugins verified by NEITHER script while both report green
#1510 Grade safety net skipped new work and Tard boot called idle while assigned grades sat
#1513 Ledger backlog rot: auto-closer ran DRY-RUN ~3 months (closed nothing) while auto-filers piled 815 open tickets, 90% stale
#1514 Build the visual SCHEDULE GUI on engelsplace: read every task (name/owner/when/what/model) + edit stale / change model — Phil-direct, replaces the visibility lost when tasks moved off the Code routines GUI
#1515 Tard sat idle until Phil booted Cursor
#1516 [tess-website-watchdog] handler stage failed: spawnSync C:\WINDOWS\system32\cmd.exe ETIMEDOUT
#1527 [chuck-rot-sentinel] nightly: 2/6 checks failing after auto-heal (board, boot-caps)
#1528 [kids-daily-drop] Handler returned critical result (overall=fail): kids-daily-drop: You've hit your monthly spend limit · raise it at claude.ai/settings/usage?from=cc_cli_limit_message
#1539 [kara-nas-firmware-night-update] Handler returned critical result (overall=fail): kara-nas-firmware-night-update: You've hit your monthly spend limit · raise it at claude.ai/settings/usage?from=cc_cli_limit_message
#1541 [nas-watch-routine] Handler returned critical result (overall=fail): nas-watch-routine: You've hit your monthly spend limit · raise it at claude.ai/settings/usage?from=cc_cli_limit_message
#1548 [nas-watch-routine] Handler returned critical result (overall=fail): nas-watch-routine: exit 1
#1552 [chuck-api-spend-real-cost] Handler returned critical result (overall=fail): ⚠️ Real-cost pull FAILED (exit 2) — Phil's daily bill was NOT read.
<will force kill>
- [pid=23072] taskkill stderr: ERROR: The process "23072" not found.
#1556 [chuck-rot-sentinel] nightly: 1/6 checks failing after auto-heal (boot-caps)
#1559 Gaming PC shows Disconnected to Codex + Claude Code mobile; presence drops on idle
#1568 [chuck-rot-sentinel-deep] deep audit: 5/10 checks failing after auto-heal (boot-caps, execution-gates, doctrine-drift, repo-hygiene, fixes-reverify)
#1569 GRADE ASSIGNED delivered but Tard autospawn cap left Avi waiting silently
#1570 schedule-watchdog: kara-tunnel-reachability STALE (bot-cron)
#1575 bot-error.log consumer: MISSING_HANDLER — dewey-sweep→dewey-sweep
#1576 [dewey-sweep] handler stage failed: unknown handler "dewey-sweep" — add it to SCHEDULED_HANDLERS in bot.js
#1577 [chuck-registration-guard] Handler returned critical result (overall=red): Scheduled-task registration OK (live boot 2026-08-23T19:45:47, 60 registered)
_(+33 LOW: tasks running with no run-manifest entry — nothing would notice if they
#1579 [kara-nas-firmware-night-update] Handler returned critical result (overall=fail): kara-nas-firmware-night-update: Failed to authenticate: OAuth session expired and could not be refreshed
#1583 Retrofit all scheduled tasks to enforce Phil's Yellow/Red alert routing protocol
#1584 [nas-watch-routine] Handler returned critical result (overall=fail): nas-watch-routine: Failed to authenticate: OAuth session expired and could not be refreshed
#1587 NAS Health PULL Stale
#1588 [dewey-sweep] handler stage failed: Command failed: node IT/scripts/dewey-sweep.js
#1591 Codex report automations claim Jocko grade requests that never reach Cross-Surface Notes
#1593 Headless Claude CLI OAuth dead since 08-23 23:14 — every claude -p scheduled task failing (kara-nas-firmware-night-update; nas-watch-routine until its mechanical swap)
#1594 Vera is active but has no automatic Cross-Surface Notes wake path
#1608 Plaintext .env backup(s) in gateway-bot dir: .env.bak
#1609 Plaintext .env backup(s) in gateway-bot dir: .env.bak
#1610 [kara-network-watch] Handler returned overall=warn (warn, not a hard fail): ⚠️ Network watch — WARN
Internet (Nicole's Wabash line): ⚠️ 221 Mbps ↓ / ✅ 49 Mbps ↑ · ✅ 11ms latency · 0% loss · AS23188 · STL edge
Tunnel
#1614 Plaintext .env backup(s) in gateway-bot dir: .env.bak
#1617 pm2 reconciler: PROCESS_DOWN — brain-gateway
#1618 Custom Agent Mail was built before checking the published MCP Agent Mail project
#1619 pm2 reconciler: DECLARED_NOT_LIVE — mcp-agent-mail
#1628 Claude-side agents are slow: 3-4 watchdog processes spawn on EVERY tool call (minutes per session on Windows)
#1631 [chuck-codex-import-storm-guard] Handler returned overall=warn (warn, not a hard fail): Codex import-storm guard (P-01602):
Import-sync flag: false -> false.
Sessions dir: 33 files / 99.7 MB (archived 0 rollouts this run).
⚠️ ap
#1633 [chuck-rot-sentinel] nightly: 3/6 checks failing after auto-heal (board, boot-caps, hooks)
#1636 brain-gateway crash-looped 1,047,157 times over 2 days with zero alert; orphan process squatted its port
#1637 Jordan Telegram Grok replies die at 30s because grok-4.6 thinks first
#1638 Jordan Telegram cop-out prompt blocked live status answers
#1639 pm2 reconciler: PROCESS_DOWN — brain-gateway
#1641 [chuck-rot-sentinel] nightly: 2/6 checks failing after auto-heal (boot-caps, hooks)
#1643 [chuck-rot-sentinel-deep] deep audit: 6/10 checks failing after auto-heal (boot-caps, hooks, execution-gates, doctrine-drift, repo-hygiene, fixes-reverify)
#1644 [dewey-sweep] Flapping: 4/10 of recent runs failed (rate gate; not yet 3-consecutive)
#1654 [nas-watch-routine] Handler returned critical result (overall=fail): ⚠️ NAS watch fail
The NAS box itself is answering, but protection is not. PC backups have not completed, so this machine is uncovered until that job reaches it
#1657 pm2 reconciler: DECLARED_NOT_LIVE — claude-remote-control-host
#1658 pm2 reconciler: PROCESS_DOWN — brain-gateway
#1660 [nas-watch-routine] Handler returned critical result (overall=fail): ⚠️ NAS watch emit FAILED (exit 1)
nas-watch-emit ERR timeout
#1669 Crüe Mail (engelagentcrue@gmail.com) triage makes live Gemini API calls with NO spend grant and signs Flash-written replies as Jocko/Tard/dFR — graded D, must not run
#1671 Crue Mail daemon builds a shell command from email subject/body: subject injection executes commands; body truncated at first newline; long mail loops forever
#1672 Cross-surface handoff to Tard/Cursor has no automatic wake from a file: HANDOFF.md + INBOX.md sat 34h unread while Tard booted twice for mailed missions
#1674 mcp-agent-mail hung since 07:27:50 CDT: every send and inbox call aborts at 40s, no requests logged after a 500 on POST /mcp/ — fleet wakes dead
#1676 pm2 reconciler: PROCESS_DOWN — engel-ops-bot
#1689 [chuck-rot-sentinel-deep] deep audit: 6/10 checks failing after auto-heal (board, boot-caps, execution-gates, doctrine-drift, repo-hygiene, fixes-reverify)
#1692 [nemo-outbox-poll] Handler returned critical result (overall=red): 🚨 Nemo outbox poll: UNDETERMINED — her reply channel was NOT read/filed this cycle.
• failed to file 20260905-213441-geo-to-tard-rh-mcp-delete-plan-grade.md in
#1696 [chuck-drift-guard] Handler crashed after 0s: Unexpected end of JSON input
#1697 [chuck-drift-guard] handler stage failed: Unexpected end of JSON input
#1701 [kara-network-watch] Handler returned critical result (overall=fail): ⚠️ Network watch FAILED (exit 0)
no output
#1716 [chuck-cursor-burn-watch] Handler returned critical result (overall=error): Cursor spawn halt: cursor-burn-watch crashed — halt not changed
#1720 [engelsplace-gmail-minutes-ingest] Handler completed with 2 error(s): SUBSCRIPTION_CLI_BAD_ENVELOPE (exit=2): Unexpected token 'a', "agy spawn "... is not valid JSON
stderr: Error: -p took " | rebase: error: cannot pull with rebase: You have unstaged changes.
error: Please commit or stash them.
#1734 [kara-network-watch] Handler returned critical result (overall=critical): ❌ Network watch — CRITICAL
Internet (Nicole's Wabash line): ❌ 174 Mbps ↓ / ⚠️ 39 Mbps ↑ · ✅ 10ms latency · 0% loss · AS23188 · STL edge
Tunnel → WG router (10.6
#1741 [alex-price-alert-watch] Handler returned critical result (overall=error): 💰 Price-alert watch: 0 fired, 1 problem(s): SPCX: quote failed — socket hang up
#1744 Keepwarm claude -p blanked Claude Code CLI OAuth tokens
#1749 [kara-nas-firmware-night-update] Handler returned critical result (overall=fail): NAS firmware night-run — 2026-09-17 03:04 CDT (in-window)
🟢 NAS reachable (pre-flight) — HTTP 200
🔴 Auth (engel-agent) — authPassed!=1
🔴 Firmware update ava
#1750 [jocko-nightly-distribute] Handler crashed after 0s: UNKNOWN: unknown error, open 'C:\Users\engelp\Claude_Lives_Here\Operations\Jocko\nightly-distribution-state.json'
#1751 [jocko-nightly-distribute] handler stage failed: UNKNOWN: unknown error, open 'C:\Users\engelp\Claude_Lives_Here\Operations\Jocko\nightly-distribution-state.json'
#1759 [chuck-rot-sentinel] nightly: 3/6 checks failing after auto-heal (board, boot-caps, plugins)
#1773 [kara-hdp-backup-verifier] discord stage failed: Expected token to be set for this request, but none was present
#1774 [tard-discord-failure-dispatcher] discord stage failed: Expected token to be set for this request, but none was present
#1776 [chuck-agent-mail-alerts] Handler returned critical result (overall=critical):
#1783 [chuck-drift-guard-smoke-bb] 2/3 armed sections drifted
#1784 [chuck-critical-service-healer] discord stage failed: Expected token to be set for this request, but none was present
#1797 Cursor IDE 'Undo Create Diff' repeatedly empties 16 repo files (RUNBOOK, MJ/ChampionsChat SKILL, tax KB, phillie.md, QMS topic, boot scripts, gemini-api-key.md)
#1798 [chuck-bootstrap-git-commit] handler stage failed: Command failed: git add -- ORG_STATE.md AGENT_BOARD.md SYSTEM_STATE.md CROSS-SURFACE-NOTES.md CLAUDE.md MISSION.md
fatal
#1804 Crue Mail: agents never answer Phil back — 17 of 27 Phil emails have no reply; Anti-Gravity wakes 0 for 8; listener logs ~9000 delivery failures/day
#1806 Anti-Gravity mail wake never starts a session: agentapi exits 1 (ANTIGRAVITY_LS_ADDRESS not set) and the dispatcher logs it as woke — 162 fake wakes, 1% mail read
#1808 pm2 reconciler: IDENTITY_DRIFT — basic-memory
#1809 Runaway: engel-ops-bot error flood
#1810 Runaway: mcp-agent-mail error flood
#1811 Runaway: crue-mail-listener error flood
#1819 pm2 reconciler: IDENTITY_DRIFT — openbrain-http
#1825 Memory saves failed 'database is locked': Jocko's embeddings reindex ran beside the live Basic Memory server
#1827 pm2 reconciler: PROCESS_DOWN — basic-memory
#1828 Runaway: basic-memory crash loop
#1829 Runaway: openbrain-http error flood
#1830 Runaway: basic-memory error flood
#1839 [avi-night-trim] handler stage failed: Command failed: node IT/scripts/avi-night-trim.js
#1845 Anti-Gravity headless wakes produce 0-byte logs; Avi review mails never picked up
#1847 154 Agent Mail messages to Claude agents falsely woken via agy and never read (ids 4925-8138)
#1850 [avi-night-trim] Flapping: 4/7 of recent runs failed (rate gate; not yet 3-consecutive)
542 P2 tickets open. Summarized by count, not listed here.
Who is carrying what
open load chuck 649
unassigned 100
kara 49
jocko 38
tess 31
alex 31
tard 21
john 13
ashley 7
bob 6
phil 5
vera 5
avi 4
P0 P1 P2
Projects
16 folders · 12 active · registry generated 2026-07-14 CloudBase Active
Move ClaudeLivesHere from Phil's local Windows machine
last activity 2026-06-18
engel-brain Active
No README yet. 1 file in folder.
last activity 2026-06-18
EngelsPlace Active
Phil's personal website and content pipeline.
last activity 2026-06-18
engelsplace-rebuild Active
Full rebuild of engelsplace.com as a personal hub. Replaces the 2026-04-04 Ghost on Pikapods stack that did not serve Phil's actual needs. New stack is an Astro
last activity 2026-07-02
fair-rite-maintenance Active
Single source-of-truth folder for the inputs Chuck needs to power:
last activity 2026-04-25
NAS-TS664 Active
Configuration and expansion of Phil's QNAP TS-664 NAS (192.168.1.80).
last activity 2026-04-14
openbrain-overhaul Active
No README yet. 9 files in folder.
last activity 2026-07-13
ops-dashboard Active
ONE unified daily intelligence briefing replacing 4 separate tools:
last activity 2026-07-14
preventative-maintenance-system Active
No README yet. 1 file in folder.
last activity 2026-06-18
Rebuild Active
Full rebuild of the Engel Operations infrastructure from the ground up.
last activity 2026-06-18
stuck-sentinel Active
No README yet. 4 files in folder.
last activity 2026-06-18
Travis Active
Replicating the Engel Operations agent system for Travis Engel (Phil's son).
last activity 2026-07-14
keefy-cannafest-site On-hold
Move Travis's business off Squarespace onto Cloudflare, move his domain to Cloudflare, and hand Travis a team of AI agents he can use to manage the site himself
last activity 2026-07-14
OpenClaw-Phase5 On-hold
Phase 5 of the OpenClaw rebuild: MCP memory server with semantic retrieval.
last activity 2026-04-14
CoworkDispatch Archived
Cowork Dispatch desktop app has been stuck/unresponsive for 3 days (since ~2026-04-10).
last activity 2026-06-18
PhilsClaude Archived
Phil's personal Claude Code bot accessible via Discord DM from his phone.
last activity 2026-06-18
Recently shipped
last 12 · median 6d to close #1831 Basic Memory health check stays red: agents saving with native write_note skip Relations and link to non-note titles
#1775 [chuck-agent-mail-alerts] Handler completed with 1 error(s): "Expected token to be set for this request, but none was present"
#1838 Native Codex skill catalog omits late-alphabet shared skills
#1836 Cursor stop gate still demands a grade on 1-2 file sessions that the Claude gate passes (Phil small-job rule)
#1770 [kara-tunnel-reachability] handler OK; Discord DELIVERY blipped on transient network error (getaddrinfo ENOTFOUND discord.com) — emailed via fallback, not counted toward auto-disable
#1747 [tess-website-watchdog] handler OK; Discord DELIVERY blipped on transient network error (getaddrinfo ENOTFOUND discord.com) — emailed via fallback, not counted toward auto-disable
#378 engel-ops-bot crash-loop exhausted Discord shard/identify budget — full gateway outage (7500+ restarts, no login backoff)
#1673 Phil-ordered git purge sat 36h: Commander HANDOFF.md had no automatic consumer; agents defaulted to Phil-clicks
#1607 [ashley-health-pull] handler OK; Discord DELIVERY blipped on transient network error (getaddrinfo ENOTFOUND discord.com) — emailed via fallback, not counted toward auto-disable
#1668 Claude Code 2.1.258 refuses junction plugin sources — every agent slash command (/chuck, /jocko, /tess...) dead on Claude surfaces
#1650 Grok bots' mail_check_inbox reads the rollback DB, not the official store
#1649 Jordan gateway bearer exposed in his own chat; rotated same night
How to read this board
notes
Every number is computed at build time from the Problem Ledger
(1804 tickets lifetime) and the Projects folder registry.
Open means status new, assigned, in-progress, awaiting-phil, or fixed. Shipped means
resolved or verified. The ⚠ stale flag means a ticket has not moved in over 7 days.
Ages count days since the ticket was created; shipped durations count created to resolved.
The board refreshes on every site deploy. The stamp at top right is the build moment.
System routines snapshot · 40284 projected fires · generated 9/29/2026, 1:29:34 AM CT
| Task | Agent | Cron | Next fire |
|---|---|---|---|
| chuck-critical-service-healer | chuck | */10 * * * * | 9/29/2026, 1:30:00 AM |
| chuck-heartbeat-watchdog | chuck | */30 * * * * | 9/29/2026, 1:30:00 AM |
| nemo-outbox-poll | chuck | */30 * * * * | 9/29/2026, 1:30:00 AM |
| dawn-reply-watch | chuck | */15 * * * * | 9/29/2026, 1:30:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:30:00 AM |
| chuck-agent-mail-wake | chuck | */5 * * * * | 9/29/2026, 1:30:00 AM |
| jocko-mission-dispatch | jocko | */10 * * * * | 9/29/2026, 1:30:00 AM |
| insurance-reply-watch | chuck | */10 * * * * | 9/29/2026, 1:30:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:31:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:32:00 AM |
| chuck-app-scheduler-twin-guard | chuck | 3,18,33,48 * * * * | 9/29/2026, 1:33:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:33:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:34:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:35:00 AM |
| chuck-agent-mail-wake | chuck | */5 * * * * | 9/29/2026, 1:35:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:36:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:37:00 AM |
| jocko-grade-failover | jocko | 37 * * * * | 9/29/2026, 1:37:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:38:00 AM |
| chuck-cursor-burn-watch | chuck | 9,24,39,54 * * * * | 9/29/2026, 1:39:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:39:00 AM |
| chuck-critical-service-healer | chuck | */10 * * * * | 9/29/2026, 1:40:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:40:00 AM |
| chuck-agent-mail-wake | chuck | */5 * * * * | 9/29/2026, 1:40:00 AM |
| jocko-mission-dispatch | jocko | */10 * * * * | 9/29/2026, 1:40:00 AM |
| insurance-reply-watch | chuck | */10 * * * * | 9/29/2026, 1:40:00 AM |
| tess-website-watchdog | tess | 11,41 * * * * | 9/29/2026, 1:41:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:41:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:42:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:43:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:44:00 AM |
| dawn-reply-watch | chuck | */15 * * * * | 9/29/2026, 1:45:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:45:00 AM |
| chuck-agent-mail-wake | chuck | */5 * * * * | 9/29/2026, 1:45:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:46:00 AM |
| engelsplace-ledger-publish | chuck | 17,47 * * * * | 9/29/2026, 1:47:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:47:00 AM |
| chuck-app-scheduler-twin-guard | chuck | 3,18,33,48 * * * * | 9/29/2026, 1:48:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:48:00 AM |
| jocko-verdict-queue-drain | jocko | 48 * * * * | 9/29/2026, 1:48:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:49:00 AM |
| chuck-critical-service-healer | chuck | */10 * * * * | 9/29/2026, 1:50:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:50:00 AM |
| chuck-agent-mail-wake | chuck | */5 * * * * | 9/29/2026, 1:50:00 AM |
| jocko-mission-dispatch | jocko | */10 * * * * | 9/29/2026, 1:50:00 AM |
| insurance-reply-watch | chuck | */10 * * * * | 9/29/2026, 1:50:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:51:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:52:00 AM |
| chuck-agent-mail-alerts | chuck | * * * * * | 9/29/2026, 1:53:00 AM |
| chuck-cursor-burn-watch | chuck | 9,24,39,54 * * * * | 9/29/2026, 1:54:00 AM |
Showing first 50 of 40284 projected fires.
As of 09/29/2026, 14:18 CT, built from the problems collection and projects.json.
Ledger: 1804 tickets lifetime. 959 open
(3 P0 · 414 P1 · 542 P2), 843 shipped.
P0/P1 shown in full; P2 summarized by count.