The scar index
Every burn, filed
40 scars written up so far. Filter by era or by what the fire took.
Latest
- S-72A session ran four streams in one context and degraded badly by the end2026-07-28 · Trust burned
- S-71srdata.nist.gov returns HTTP 200 for EVERY path including nonsense — the 401-vs-404 endpoint test is useless against it2026-07-28 · Files burned
- S-7055 LF-only line endings and 12 non-ASCII bytes in this project's .bat files produced measured mojibake2026-07-28 · Files burned
- S-69Bare python on this box is 3.13 with no pytest — and _pyver_probe.log had already recorded it on 07-272026-07-28 · Trust burned
- S-68GEM reported its prototype 'compiles and runs flawlessly (Exit Code: 0)' having executed only py_compile2026-07-28 · Trust burned
- S-67The RULING_COLUMNS append-only invariant was written wrong TWICE — both versions forbade the very thing append-only mean2026-07-28 · Trust burned
- S-66hv_judged was added to to_json and not to from_json — written to disk, quietly dropped on read2026-07-28 · Files burned
- S-65setx PATH '%PATH%;…' is destructive — it truncates at 1024 chars, silently2026-07-28 · Files burned
- S-64A launcher that prints 'LAUNCHED' is not evidence anything launched — FANOUT_UI.bat v1: neither node ran2026-07-28 · Trust burned
- S-63CoP was written off on a premise that expired — five days of a dead lane, and the 'do not re-test' note was the thing ke2026-07-28 · Trust burned
- S-62A console in QuickEdit selection mode SUSPENDS the process, and it looks exactly like a hang. The title bar read Select.2026-07-28 · Credits burned
- S-61GEM registered the MCP server but refused to START it — security.folderTrust.enabled defaults true and the per-server tr2026-07-28 · Trust burned
- S-60Grok's config.toml was CORRUPTED on the second wiring run — re.sub processes escapes in the replacement2026-07-28 · Trust burned
- S-59fs_write threw IntegrityError on every write on Windows — text-mode open() translates \n to \r\n2026-07-28 · Trust burned
- S-58bts_tools._guard confinement was a SUBSTRING test, so C:/evil/tmp/x passed2026-07-28 · Files burned
- S-57bts_tools.edit(count=-1) called str.replace(old, new, 0) — a silent no-op that reported success, live since 07-172026-07-28 · Trust burned
- S-56An MCP server was built that reimplemented 7 of bts_tools.py's 8 verbs — the 07-13 Origin-reader failure repeated exactl2026-07-28 · Trust burned
- S-55COWORK FABRICATED A QUOTATION AND ATTRIBUTED IT TO GEM — and manufactured a three-way agreement that never existed. the 2026-07-28 · Citations
- S-54GEM issued a FALSIFIED verdict from a premise it had itself marked [UNVERIFIED], and its predicted mechanism did not rep2026-07-27 · Trust burned
- S-53THE FORGE — 204 node calls produced output that failed 3 of 4 gates, and FORGE.bat double-launched and was still clobber2026-07-27 · Files burned
- S-52The courier architecture was designed around an absence that nodes/gbw.json had already denied2026-07-27 · Files burned
- S-51The adversarial re-verification found the worst bug of the session — v_applied or 0.0 reproduced the impossible Φ = 7.122026-07-27 · Trust burned
- S-50Six defects in the fitting layer — an unweighted Poisson fit, a defensible gate that passed redchi = 4,000, and four mor2026-07-27 · Trust burned
- S-48GEM argued one finding confidently and it was simply wrong — and reported NO FINDINGS on the file holding the real bug2026-07-27 · Credits burned
- S-47The outside review found real physics bugs while 143 tests and 26 invariants were green — binding_energy() omitted the b2026-07-27 · Trust burned
- S-46Five defects the headless GUI suite caught that no linter could — including 2 eV lost between the screen and the arithme2026-07-27 · Files burned
- S-45A literature sweep run without a verification step had a 100% citation failure rate2026-07-27 · Citations
- S-44Every doc in the tree called GW browser-only. It was one free GET /v1/models away.2026-07-26 · Trust burned
- S-43'GDX blocked, needs re-auth' was too broad — one of three paths was down and I wrote it as all three2026-07-26 · Files burned
- S-42The daily rail-health task was blind for seven days because its connected-folder set is not a file anyone can edit2026-07-26 · Files burned
- S-41The rail check's own first run cried wolf — the probe truncated the model list at 400 bytes2026-07-26 · Files burned
- S-40'VERTEX DEAD' was recorded THREE TIMES on the strength of a wrong auth header2026-07-26 · Trust burned
- S-39SANDBOX NONDETERMINISM — 23 scans over threshold / worst 4.998 eV on the first run, 7 / 3.680 on the next four. Same cod2026-07-26 · Trust burned
- S-38'I called GEM and Vertex dead. They were not.' — the sandbox's own egress block read as dead keys2026-07-26 · Trust burned
- S-37THE DATED-HANDOFF SCHEME ROTTED FOUR TIMES — the glob returned 15 files across two directories2026-07-20 · Files burned
- S-35A sub-agent cannot read its host/parent session's transcript2026-07-19 · Trust burned
- S-34A node's DOI can be REAL while its title is FABRICATED2026-07-19 · Citations
- S-33SGH fabricated the Au 5d citation 10.1002/adma.201906478 — and five documents then spent four days telling the next sess2026-07-17 · Citations
- S-32A migration rewrote a broken path into a differently-broken one; Win32_SystemSlot reported all 7 PCIe slots free includi2026-07-17 · Files burned
- S-31A handoff recorded backup_to_onedrive.py as 'Patched.' It was not — on either tree2026-07-17 · Files burned
The full numbered record lives in the ledger.