Field notebook · No. 0716
Solarolabs
This notebook records measured misses, not a catalog of tools.
BTS_MESH is the software. KMesh is the instance that was live. The charter's first rule, instantiated July 16: the mesh is infrastructure for the dissertation, not the point of it. This page is one note for every day from July 1 through October 29. The burn is August 17, 2026, when a delete command cleared the research tree. A day with no file in the record says so.
Wednesday
No scar is filed on this day. Fifteen days later, S-25 records that the public mirror was still serving the July 1 folder-access file, and every check returned 200. A publish is not a sync. That is a reason the mesh keeps a host copy and does not trust a green light.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
The operator named the routines BootUP and TidyUP, and ordered that a new class of item be added to the checklist when it appears. The mesh is those lists, kept in files, not in a session's memory.
Saturday
The boot note points the next session at a primer dated this day, then at the task list. The mesh starts from a file. It does not start from what the last session meant to do.
Sunday
S-01. Three Chrome-bridge denials were the operator being away, not the operator refusing. If the operator is away and it's denied, note it and queue bridge work; do NOT retry-spam (3 denials on 2026-07-05 = the operator away, not refusal). The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated.
Monday
S-02. The ROLD commands file was lost to disk truncation and had to be reconstructed. ⚠ RESTORED 2026-07-06 after disk truncation (see .bak_2026-07-06- files): reconstructed from Doping-session context; this IS the canonical current version. Post-edit integrity rule: wc -c + tail-check after every edit of this file. A file that ends early is why a write is measured before it is trusted. 1 other scar is dated this day. This is the one that states the reason.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
S-04. GDX write silently failed on stale connector auth, and the node said it could write. 2026-07-10/11 (mesh/dashboard + GDX session): GDX write-requires-reauth class — Grok's Drive write silently failed on stale connector auth (SGH walked back an earlier "yes I can"); the operator re-auth fixed it and the round-trip proved out. Lesson: don't trust an agent's self-reported connector capability — v The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated.
Saturday
S-04. GDX write silently failed on stale connector auth, and the node said it could write. 2026-07-10/11 (mesh/dashboard + GDX session): GDX write-requires-reauth class — Grok's Drive write silently failed on stale connector auth (SGH walked back an earlier "yes I can"); the operator re-auth fixed it and the round-trip proved out. Lesson: don't trust an agent's self-reported connector capability — v The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated.
Sunday
S-05. MOUNT HIT: jack_command.html served pinned at a truncated 63,824 bytes. 1. TRUNCATION (known): jack_command.html served pinned at 63,824 bytes (host: 1,146 lines) and A mount that lies is why the mesh checks the host disk, not the view a tool was handed. 6 other scars are dated this day. This is the one that states the reason.
Monday
S-14. The Ch4-review publish stalled because [path] was not in the mount list. Without this mount Cowork cannot publish, cannot read the bat, and has to hand the operator a RUN_ME_.bat to double-click — which is exactly what stalled the Ch4-review publish on 2026-07-13. A mount that lies is why the mesh checks the host disk, not the view a tool was handed. 2 other scars are dated this day. This is the one that states the reason.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
S-16. GDX mounts for Claude's file tools but NOT into the bash sandbox. 🔴 ALSO 2026-07-15 — [path] Drive\BTS_SGH_Handoff (GDX) mounts for Claude's file tools but NOT into A mount that lies is why the mesh checks the host disk, not the view a tool was handed. 7 other scars are dated this day. This is the one that states the reason.
Thursday
The charter is instantiated. BTS_MESH is the software. KMesh is the live instance, infrastructure for the dissertation, not the point of it. Chapters outrank the mesh. The same day the ledger records S-22: TidyUP2 produced two false alarms out of four checks.
Friday
S-29. MOUNT HIT #12 — a FALSE SyntaxError at line 106 nearly ate bts_paths.py, the resolver everything depends on. #12 — a FALSE SyntaxError in bts_paths.py, the resolver everything depends on. The sandbox reported A mount that lies is why the mesh checks the host disk, not the view a tool was handed. 5 other scars are dated this day. This is the one that states the reason.
Saturday
The session log for this day records: 🔴 READ THIS FIRST — THIS TASK'S PREVIOUS PROMPT WAS FALSE AND PRODUCED A FABRICATED ALARM.
Sunday
S-34. A node's DOI can be REAL while its title is FABRICATED. 1. 🔴 A NODE'S DOI CAN BE REAL WHILE ITS TITLE IS FABRICATED. GEM/SGH handed back 10.1103/PhysRevB.20.4126 An unchecked citation is why a second node exists, and why two copies of one model do not count as two. 2 other scars are dated this day. This is the one that states the reason.
Monday
S-37. THE DATED-HANDOFF SCHEME ROTTED FOUR TIMES — the glob returned 15 files across two directories. Why: the dated scheme rotted FOUR times (12c→12d, 12d→07-15, →07-17, and on 2026-07-20 this line The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
The session log for this day records: 🔴 READ THIS FIRST — THIS TASK'S PREVIOUS PROMPT WAS FALSE AND PRODUCED A FABRICATED ALARM.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
The session log for this day records: 🔴 READ THIS FIRST — THIS TASK'S PREVIOUS PROMPT WAS FALSE AND PRODUCED A FABRICATED ALARM.
Sunday
S-38. "I called GEM and Vertex dead. They were not." — the sandbox's own egress block read as dead keys. 1. I called GEM and Vertex dead. They were not. Every googleapis.com host returned HTML 403 A green status with nothing behind it is why the mesh reads the artifact, not the exit code. 6 other scars are dated this day. This is the one that states the reason.
Monday
S-47. The outside review found real physics bugs while 143 tests and 26 invariants were green — binding_energy() omitted the bias. 1. binding_energy() OMITTED THE BIAS. BE = (hv − E_an) − KE, missing + V. A biased sample A green status with nothing behind it is why the mesh reads the artifact, not the exit code. 9 other scars are dated this day. This is the one that states the reason.
Tuesday
S-59. fs_write threw IntegrityError on every write on Windows — text-mode open() translates \n to \r\n. THE SANDBOX SUITE WAS GREEN AND WINDOWS STILL FOUND THREE BUGS. All three were OS-shaped. A green status with nothing behind it is why the mesh reads the artifact, not the exit code. 17 other scars are dated this day. This is the one that states the reason.
Wednesday
S-74. The R2 publish took 16 minutes to transfer ZERO bytes — 87 rclone invocations, 970.3 s. 🔴 AND THE LOG ANSWERED "WHY IS R2 SLOW" WITH A NUMBER: 16 MINUTES TO TRANSFER ZERO BYTES. A public copy is not the bus. The mesh keeps those apart. 5 other scars are dated this day. This is the one that states the reason.
Thursday
S-89. "the operator signed in to CoP on 2026-07-29" was recorded as fact and is FALSE for the CLI. [path] states: "CoP Copilot CLI 1.0.75 — ✓ registered … the operator signed in 2026-07-29 on The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated. 16 other scars are dated this day. This is the one that states the reason.
Friday
S-106. THE PUBLISH WROTE ITS LOG INTO THE TREE IT WAS PUBLISHING, AND ONE FAILED FILE SILENTLY DISABLED THE ENTIRE DELETE PASS. Measured across THREE consecutive publishes tonight. Same error, same file, every time: A public copy is not the bus. The mesh keeps those apart. 10 other scars are dated this day. This is the one that states the reason.
Saturday
S-108. THE HANDOFF'S OWN CONSTANTS WENT STALE INSIDE THE TIDYUP THAT WROTE THEM. [path] stated verify_pointers 129 and scars 106. TidyUP2 CHECK 4 measured 130 and 107. The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated. 1 other scar is dated this day. This is the one that states the reason.
Sunday
S-113. COWORK MISIDENTIFIED THE DEAD DRIVE TWICE IN ONE NIGHT, BOTH TIMES BY REACHING A CONCLUSION AHEAD OF THE IDENTIFIER. First: advised a head-stack swap and a donor ST8000NM0055 from the drop history alone — A green status with nothing behind it is why the mesh reads the artifact, not the exit code. 6 other scars are dated this day. This is the one that states the reason.
Monday
S-120. THE PRE-RUN ESTIMATE WAS PRINTED AS THE MEASURED COST, IN THREE DOCUMENTS, WITH THE LEDGER SITTING RIGHT THERE. The RFA REVIEW.bat menu said "approx $1.70" — a legitimate estimate, written before the call, The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated. 11 other scars are dated this day. This is the one that states the reason.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
S-130. FIVE OF SEVEN CITATIONS FROM ONE RAIL WERE FABRICATED — AND BOTH LOAD-BEARING ONES WERE INVENTED. the operator: "Did you delegate this? See what the CREW can find as well." The same Oklahoma An unchecked citation is why a second node exists, and why two copies of one model do not count as two. 2 other scars are dated this day. This is the one that states the reason.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
S-134. THE MONITORS WORKED WHENEVER ANYONE WAS WATCHING THEM. FOUR TOOLS DIED ON AN EMOJI, AND TWO OF THEM EXITED 0 WHILE DOING IT.. rail_check.py, bts_drive_health.py, bench_drive.py and bts_kdash_feed.py all print 🔴 and An unchecked citation is why a second node exists, and why two copies of one model do not count as two. 6 other scars are dated this day. This is the one that states the reason.
Wednesday
S-140. "GREEN ×4 — THE FIRST REAL DRIVE READING THIS MESH HAS EVER HAD" WAS GREEN ON Nones. bts_drive_health scored uncorrected error counters with if isinstance(v, int) and v > 0: RED. A green status with nothing behind it is why the mesh reads the artifact, not the exit code. 6 other scars are dated this day. This is the one that states the reason.
Thursday
S-145. S-134 INVERTED: THE TOOL WORKED IN THE HARNESS AND DIED IN FRONT OF THE HUMAN — AND IT DIED ON THE LINE NAMING THE CAUSE. runner_doctor.py was written on 2026-08-12 to diagnose the queue runner. the operator ran it from The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated. 1 other scar is dated this day. This is the one that states the reason.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
The operator ordered the roadmap: rails running, then phone and voice, then share with the next peer. A phone is a client that is not this box, and so is a peer. The charter says the mesh of meshes is not ready, and that this is the correct report. The ledger the same day records S-147: THE NEGATIVE CONTROL SHARED THE POSITIVE CHECK'S DEFECT, SO IT CERTIFIED NOTHING — AND IT PRINTED A ✅ SAYING SO.
Sunday
S-149. The mailbox was correct syntax on the wrong surface, so a dead channel was indistinguishable from a quiet one. The same defect as S-148 in different clothing: correct syntax, wrong surface — letters written A green status with nothing behind it is why the mesh reads the artifact, not the exit code. 1 other scar is dated this day. This is the one that states the reason.
Monday
This is the day it burned down. A remove-directory command deleted every subdirectory of the research tree. The recovery plan, written the next morning, says the bat was the mesh's own. The case file on the other drive was not in the path. The public copy of the mesh survived because publish copied and did not sync, except for two folders.
Tuesday
The day after the burn. The session-log archive had survived one level up, and a replay rebuilt files the sweep had classed as a total loss. The same day the ledger records a standing delete order aimed at the disk that now held the only full copy. A read-hazard had been written as a destroy order.
Wednesday
The bytes were found and staged. Nothing put them back. For three days the boot file kept pointing at a chapter draft that was not on the live tree. Recovering a file and restoring it are different acts.
Thursday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Friday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Saturday
The restore was counted. The working folder went from 13 files to 1,216. A restore is not done until a count and a size say so. An archive that had sat unopened held 1,045 of the files called lost.
Sunday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Monday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Tuesday
S-155. CLASSES DISCOVERED (TidyUP2 check 10) — two evidence-handling classes from the legal stream. (a) Self-maintained business listings as party-statement evidence. A Google Business Profile the The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Friday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Saturday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Sunday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Monday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Tuesday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Wednesday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Thursday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Friday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Saturday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Sunday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Monday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Tuesday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Wednesday
The morning note still carries open mesh work, and no separate scar is filed. - M-06 — COSMOS build at [path] The mesh exists so that item stays on a page instead of in a session.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
A fat session on the flash model hit resource exhausted and was stored as a dead error instead of a queue. A 429 is a wait, not a corpse. The mesh exists so a rate limit does not end the work.
Wednesday
S-156. OpenRouter HERO 429 was the upstream free pool, not our weekly cap — and several "fails" were other classes. Measured while seating CCrew by TPS. GET /auth/key: is_free_tier=false, weekly remaining ~149.8/150. Gemma :free and Qwen :free HTTP 429 body: Provider returned error, limit_source=upstream_provider_shared_pool (Google AI Studio / ModelRun). Retrying those slugs does not spend our cap; it hits a shared free pool. The charter's first rule is why this is written down: the mesh is infrastructure for the dissertation, and a miss that is not recorded is repeated.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Friday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Saturday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Sunday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Monday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Tuesday
Eight shifts put an existing scar in the lead, one each half hour, and added none. The page changed. The record did not grow.
Wednesday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.
Thursday
No measured scar in the ledger is dated this day, and no mesh file in the record carries this date. The page does not invent one.