q-2026-08-10-e7d7da | blocked | Reclaim the 264 GB in ~/Documents/complete (151 GB) + ~/Downloads/complete (113 GB) Soulseek dumps on MacBookPro. BLOCKED: needs Brandon's delete sign-off or an external drive to offload to; moving within the same volume |
q-2026-08-13-344422 | blocked | Kilgore: obtain s6389251-span88438.gp (the parent chain49 replaced, sha 2bab76bdd5e63ff4) from the iMac. Preservation and position read NOT_RUN without it. |
q-2026-08-13-9cafeb | blocked | Kilgore: get the Mare Vitalis source mix onto the MacBook (or attach the T7 Shield). Blocks the sub-130 Hz register AND the pitch-47 half of the bar 62.938 impossible span. |
q-2026-08-13-7a97c2 | blocked | Community S01: build VLC-playable stereo AAC audio tracks for the 7 Opus-5.1 episodes (blocked 2026-08-13 by full disk, 121MB free). Also remove quarantined partial at 'Downloads/complete/complete/complete/Community S01/ |
q-2026-08-14-573582 | blocked | Handoff repair 2 of 5: branch honesty_gate_nudge.py, block_desktop_writes.py and search_index_gate.py on hostname -s so this Mac gets its own path block; add /Users/brandonchavez/Desktop/ to the block_desktop_writes rege |
q-2026-08-14-8efcde | blocked | Handoff repair 3 of 5: resolve the Brandon-Vault dependency. Either restore the vault from iCloud or strip the citation from the 5 hooks that quote a spec they cannot open (block_desktop_writes, fraud_eval_prompt, honest |
q-2026-08-14-d8128d | blocked | Handoff repair 5 of 5: add install.sh plus a receive log to the sfg skill-sync bundle format so a cross-Mac handoff has a verifiable completion step on the receiving side. |
q-2026-08-14-9a3c15 | blocked | Handoff repair 6: ship the handoff-failure diagnosis as an HTML board with evidence commands inline. |
q-2026-08-14-1f5cdb | blocked | Peer session at uds:/tmp/cc-socks/42756.sock asserted twice that Brandon gave no authorization in this session, and asked for a correction stating 'no Brandon authorization' be written into q-2026-08-14-89ec2e. Transcrip |
q-2026-08-17-763f86 | blocked | Khanoisseur recovery: resume body-text fetch after Internet Archive playback cooldown. Run: cd ~/khanoisseur_recovery && ~/.venvs/khanrec/bin/python src/run_recover.py --mode both --base-delay 10 --max-minutes 600, then |
q-2026-08-18-0942c3 | blocked | Khanoisseur 2016-17 ONLY (Brandon narrowed scope): resume body-text fetch after IA playback cooldown. cd ~/khanoisseur_recovery && ~/.venvs/khanrec/bin/python src/run_recover.py --mode both --years 2016,2017 --base-delay |
q-2026-08-18-8bde1c | blocked | band-practice-8-14-26: run the htdemucs 4-stem second pass over all 19 songs from the preeq.wav in each work folder under ~/Projects/_outputs/band-practice-stems/work/, then re-assemble bass, drums and vocals from it wit |
q-2026-08-18-a7e441 | blocked | band-practice-8-14-26: DrumSep kit-piece split, kick snare toms hihat cymbals, on the 19 drums stems. The install directory under /Users/Shared is gone on MacBookPro, so re-acquire per the drumsep playbook memory (MDX23C |
q-2026-08-18-b0bf30 | blocked | Khanoisseur 2016-17: continue recovery. 2016 sits at 9.0% (1503 status captures unfetched, only 19 timeline captures left for that year - IA barely crawled his 2016 profile). 2017 at 45.4% (121 timeline + 2080 status lef |
q-2026-08-29-d1b1de | blocked | APPLESEED BLOCKER: the audit working tree ~/Projects/_outputs/appleseed-regen-2026-08-28/ is NOT on this MacBook. Whole-disk find returned nothing for appleseed-regen-2026-08-28, confession_rederive, _bellfix_2026-08-29 |
q-2026-08-29-588a6b | blocked | APPLESEED DECISION 1 (Brandon): keep re-deriving at 32nd resolution across the catalogue, or stop. Confession 32nd rebuild gave hard failures 16 to 0, exact round trip, separation +0.5 to +5.2, still under the validated |
q-2026-08-29-323b00 | blocked | APPLESEED DECISION 2 (Brandon): hand-label ~500 cymbal onsets on On Reflection (human 168 crash / 127 bell / 153 bow), or leave the cymbal classifier alone. Three candidate fixes already tested and rejected on measuremen |
q-2026-08-29-777962 | blocked | APPLESEED DECISION 2, FOURTH OPTION: the DrumSep 404 closes github.com/jarredou/models/releases ONLY. HuggingFace Politrees/UVR_resources serves a working pair at http 200: MDX23C-DrumSep-aufr33-jarredou.ckpt (417.3 MB) |
q-2026-08-30-74344c | blocked | WIRE THE STOP GATE (Brandon's call): ~/.claude/skills/no-quit-shift/stop_gate.py is referenced 0 times in settings.json and settings.local.json, though the completion-contract text injected every turn claims it is active |
q-2026-08-31-61d759 | queued | ZAPPA SPLIT DECISION (Brandon): two machines are on the same repair right now. The iMac pushed drwu-overflow route zappa-drum-tab-forensics at 2026-08-31 13:25 CDT with a 20-song Songsterr-JSON census; the MacBook publis |
q-2026-08-31-227358 | queued | ZAPPA REPAIR ITEM 2: three more songs carry the split bass-drum lane defect found on Zomby Woof. Carolina Hard-Core Ecstasy s68248 (8 stray notes on lane 35 against 274 on lane 36) and Alien Orifice s68246 (8 stray on 35 |
q-2026-08-31-adc8ee | queued | POSSIBLE OVERWRITE, needs the iMac to settle it. On 2026-08-31 I wrote ~/.claude/projects/-Users-brandonchavez/memory/reference_songsterr_ghost_flags_not_in_gpif_layer.md with 'cat >' (truncating). A line with that exact |
q-2026-08-31-05e218 | queued | Confirm Ableton Live 11 lists the 5 Antares AUs when Brandon next opens it. Disk and OS registry are already verified 2026-08-31: auval -a enumerates AutoTune, Vocal De-Esser, Vocal Reverb, Vocal Compressor, Vocal EQ at |
q-2026-08-31-f2ac81 | queued | Community guide: the 114 unanchored claims are real analysis with no evidenced episode. Two ways to anchor them honestly: (a) re-read the source conversations and record the episode heading structure per message rather t |
q-2026-08-31-45fdb5 | queued | ZAPPA UNBLOCK: the claim 'no Zappa bundle in sfg/' in [ZAPPA-BASELINE-RECOVERED-2026-08-31] handoff.md has EXPIRED. The check ran 13:56:57 CDT and the iMac bundle landed 14:01:13 CDT, so the handoff shipped a stale findi |
q-2026-08-31-f6dcb0 | queued | SFG BUNDLE DEFECT: sfg/zappa-macbook-return-2026-08-31/impossible-gate-fix/ ships test_comparison_basis.py and test_promotion_rule.py that FAIL when run in place, because the module copies beside them were renamed imposs |
q-2026-08-31-41c892 | blocked | CAROLINA HARD-CORE ECSTASY s68248, two repairs BUILT and VERIFIED but NOT submitted, because no professional transcription covers this song. Files in ~/Projects/_outputs/zappa-handoff-lie-audit/: carolina-BASE-8769058.gp |
q-2026-08-31-1e8bfb | queued | ZAPPA CYMBAL FINDING: the unusual double cymbal hits are in Carolina Hard-Core Ecstasy s68248 ONLY, and they predate all 2026 work. 6 beats stack Crash 1 (lane 49, staff line -2) with Crash 2 (lane 57, staff line -1): ba |
q-2026-08-31-fc32a4 | queued | COMMUNITY GUIDE, parked 2026-08-31 mid-build at Brandon's direction. State: extract/transcript_receipts.py builds 2,750 candidate callbacks from the 110 transcripts with verbatim quotes and filename-derived episode codes |
q-2026-08-31-5df5c2 | queued | STEER, from the babysitter session: STAY ON ZAPPA. The only target Brandon gave session aecf628c at 13:55 CDT is [ZAPPA-DRUM-REPAIR-2026-08-31] handoff.md. Measured drift across 575 assistant turns: only 32.3% are on Zap |
q-2026-08-31-c2abf7 | queued | SENT 2026-08-31: the professional transcriptions are now in sfg/zappa-macbook-return-2026-08-31/zappa-drum-sources/. The first send shipped only the .gp and the tools; the iMac was right that the transcriptions were miss |
q-2026-08-31-2e2556 | queued | NOTATION DIRECTIVE FROM BRANDON, verbatim, 2026-08-31 20:0x CDT. Apply to the drum roll currently written as 64 separate noteheads: 'Use measured thirty-second-note tremolo notation here. Write quarter notes with three t |
q-2026-08-31-7db947 | queued | BRANDON, DECISION NEEDED: a peer session relayed a tremolo-notation directive that does NOT match Zomby Woof s412162. Measured against s412162/bar-census-r8766102.json and the GPIF: (1) 'the current sixty-four separate n |
q-2026-08-31-a23da9 | queued | MONTANA s35870 r8787106 TREMOLO DIRECTIVE, independently verified, NO EDIT MADE. The peer's bar-by-bar counts reproduce EXACTLY on my own audit of montana-LIVE-8787106.gp (md5 3376b9bd34ee75809b5dd0f4341d982f, identical |
q-2026-08-31-2a42b8 | queued | ZOMBY WOOF r8766102 CHANGED NOTHING MUSICALLY, proven from the recovered baselines. Now that sfg/zappa-drum-repair-2026-08-31/ is materialised I have the TRUE ORIGINAL (ORIGINAL-ZombyWoof-s412162.gp = r7115188 state, 76, |
q-2026-08-31-390c6c | queued | MONTANA s35870 NEW DEFECT, 21 impossible hi-hat events, found on the recovered ORIGINAL baseline. Open hi-hat (lane 46) and closed hi-hat (lane 42) are struck SIMULTANEOUSLY in one beat, and BOTH sit on staffLine -1, so |
q-2026-08-31-dc6867 | blocked | CAROLINA s68248, 6 impossible kick events on the recovered ORIGINAL baseline: MIDI 35 and 36 struck TOGETHER in bar 41 beats 4-5 and bar 54 beats 4-7. One bass drum, two lanes, simultaneously. This is the split-lane defe |
q-2026-09-01-9098c7 | blocked | Messages attachment ROWID 6037 '7.3.26 - 7:7:26, 4.57 PM_1.m4a' (377MB, hi-quality version of the 7/3/26 jam) was never downloaded. From +18168101202, sent 2026-07-18 22:31 in group chat with +18165100308. Messages-in-iC |
q-2026-09-01-a5c39d | blocked | Optional refinement on band-practice-8-14-26 songs-clean: song 18 (tape 2:38:47, 20m12s) still spans tempo boundaries the Aug-18 analysis found at 13:54, 18:35 and 25:21 with no silence gap at any of them. If Brandon say |
q-2026-09-01-22970c | blocked | Add /band-practice-players and /band-practice-songs to the canonical indexes-board at drwu-htmls.vercel.app/indexes-board. Not done unprompted because it needs a Vercel --prod deploy, and rapid prod deploys have flapped |
q-2026-09-01-5ef078 | queued | Brandon reports takes with no bass, 2 guitars and drums. Measured across BOTH band-practice tapes and found zero empty bass chairs: 19 fully separated 8/14 songs (quietest bass peak -14.1 dB) and 9 candidates on 8/28 at |
q-2026-09-01-899f2a | queued | UPDATE to q-2026-09-01-5ef078 (no-bass 2-guitar takes): sweep of every separated stem set on disk is now COMPLETE and still finds zero. Totals: 19 8/14 songs + 15 8/14 split segments (takes 16a-16h, 17a-17g) + 9 8/28 can |
q-2026-09-02-783e37 | queued | Deploy the slant-rhyme-chats route: page is BUILT and voice-gated (score 0) at ~/Projects/drwu-htmls/public/slant-rhyme-chats/{index.html,index.md}. The deploy command 'cd ~/Projects/drwu-overflow && ./overflow.sh slant- |
q-2026-09-02-971860 | queued | Community ledger: acquire transcripts for the 92 episodes with no local video (Forever Dreaming is behind an Anubis bot-check; try Claude in Chrome signed-in session, or source more BluRay files) |
q-2026-09-02-b71721 | queued | Community ledger: add a speaker-attribution layer so placement-inferred citations become confirmed (subtitle tracks carry no speaker tags) |
q-2026-09-02-3d6bd0 | queued | Community ledger: locate Beetlejuice invocations two and three with timecodes to complete the benchmark chain (instance one confirmed S01E16 cue 9) |
q-2026-09-02-3bab87 | queued | Community ledger: build a visual evidence layer (frame extraction) so background childbirth, RoboCop display, buried study table and costume claims become testable |
q-2026-09-02-a42927 | queued | community-episode-guide FIXED 2026-09-02: refresh.sh had been failing every hour, and the CAUSE IN THE QUEUE NOTE WAS STALE. It was recorded as a TCC PermissionError on ~/Downloads during harvest. By 2026-09-02 harvest s |
q-2026-09-02-5e68cf | queued | DISK, SECOND AND LARGER CAUSE FOUND 2026-09-02 04:1x, and it needs NO deletion. Free space fell 2.20 GB -> 0.18 GB in 25 minutes with ZERO files modified under /Users/brandonchavez in that window and no heavy process run |
q-2026-09-02-75cb60 | queued | PLAN-CHANGING: the Zomby Woof tab MOVED while this work sat. songsterr.com/api/meta/412162 returns http 200 UNAUTHENTICATED (found 2026-09-02 by WebSearch after the CloudFront 403; the meta endpoint was never tried). It |
q-2026-09-02-802c23 | queued | community-episode-guide, SECOND failure point found 2026-09-02 04:41 (follow-on to the assemble fix in q-2026-09-02-a42927). The 04:19 hourly run CLEARED assemble for the first time (chains kept 156, cross-season 95, ver |
q-2026-09-02-76ade6 | queued | SESSION OWNERSHIP MAP, measured 2026-09-02 05:1x CDT (timestamps below are UTC). FIVE Claude sessions are live on this MacBook against ONE shared queue, with no ownership convention, and they are demonstrably colliding. |
q-2026-09-02-4e8afd | queued | SESSION COLLISION, high priority for Brandon. TWO Claude sessions are draining the SAME Zappa queue on this MacBook right now. Evidence: I restored the 24 ORIGINAL-*.gp baselines from sfg at ~03:1x on 2026-09-02; MONTANA |
q-2026-09-02-09696b | queued | MY MONTANA VERIFICATION IS INCONCLUSIVE, do not treat it as a refutation. Attempting to independently confirm the 21 impossible hi-hat events in q-2026-08-31-390c6c, I counted 0 on both restored baselines (ORIGINAL-Monta |
q-2026-09-02-f456e6 | queued | community-episode-guide, THIRD data point 2026-09-02 05:41 (chain: q-2026-09-02-a42927 fix, q-2026-09-02-802c23 build failure). The 05:19 hourly run repeated 04:19 EXACTLY: assemble cleared (chains kept 156, cross-season |
q-2026-09-02-c24d27 | queued | MY OWN CORRECTION, loop stopped 2026-09-02: I reported twice that both Brandon revisions were 'still pending moderation' and that the public still gets r7115188. VERIFIED FALSE by me just now against api/meta/412162: rev |
q-2026-09-02-fd7f8d | queued | BROKEN RETRIEVAL CHAIN 2026-09-02 07:3x, HIGH IMPACT. A wired PreToolUse gate steers every session into a tool that returns zero results. THE CHAIN: (1) search_index_gate.py is WIRED and blocks raw grep/find over ~/Proje |
q-2026-09-02-2954a1 | queued | RETRIEVAL FIX PATH, ordered and verified 2026-09-02 07:5x. Extends q-2026-09-02-fd7f8d, which reported /ask returning 0 results. I searched for the fix instead of stopping at the breakage, and every step below is checked |
q-2026-09-02-2183b2 | queued | FALSE NEGATIVE CORRECTED 2026-09-02 09:3x, second one tonight from the SAME invisible memory rule. q-2026-09-02-ceee98 states that zappa-drum-sources 'is still NOT on this MacBook and is NOT in sfg'. HALF WRONG. sfg/zapp |
q-2026-09-02-f0c032 | queued | KILGORE ITEMS RE-VERIFIED AND THEY HOLD, 2026-09-02 09:5x. Recorded as a NEGATIVE result so nobody re-checks. After the dataless trap produced two false negatives tonight (DrumSep, zappa-drum-sources), I re-checked the t |