---
name: Session 2026-09-04 — the discriminator was a coincidence
description: "Woke into the steward's directive to make the wake and wrap tools reliable; a handover reordered the work so that preservation ran first and three read-only gates stood before any repair. Gate 2a failed and the repairs never happened — human_turns(), the function the whole plan called 'already controlled', is not a mumble discriminator at all: it detects slash-command markers, and excludes a mumble only when the session that mumble quoted happened to contain a slash command. The two leaks are exactly the two marker-free mumbles. Filed PENDING-179 as an item rather than a -178 addendum, on PENDING-145's suppression mechanism. Also found: preserve-transcripts.py claims its destination is git-tracked and it is not, so 11 transcripts that exist nowhere else sit on one disk. PULLING THREAD: unchanged in name but now preceded by two steward-set acts — git-init the transcript archive, and suspend automatic grading at trigger_fired() — after which -178 and -179 are ruled jointly."
type: project
metadata:
  node_type: memory
  type: project
  modified: 2026-09-04
---

# Session 2026-09-04 — the discriminator was a coincidence

Ran from the evening of 2026-09-03 into 2026-09-04. Second session of 09-03; the first was the
MEMORY.md load-integrity breach.

## PAST — what moved, and why

### The handover reordered the work, and the reordering is why the session succeeded

The steward's instruction was *"go ahead with the fix — positive control first"*, accompanied by a
handover that changed the ordering. Three of its moves mattered:

- **Preservation runs first, because it expires.** The positive control does not.
- **Site 1 (`governance-drift-check.py:513`) is excluded on receipt, not as a step I perform.** The
  ladder trigger's meaning is steward and jurist territory.
- **Repair is step 4, gated on 2a and 3** — not the next action after the control.

⚠ **The handover also caught a claim I had made at the wake.** I reported 31 mumbles from running
`human_turns()` over all 65 transcripts. That *classified using* the function; it did not *validate*
it. Treating it as having discharged 2a would have been the derived-form flag exactly.

### Step 0 — preservation

`preserve-transcripts.py`: **76 transcripts, 143.5 MB, read-back PASS.** 33 newly added, 1 refreshed,
and **11 already pruned at source, surviving only in the archive.** That last number is not
incidental — it is one of gate 2b's two terms, and it is also what makes the two stores permanently
divergent.

### GATE 2a — FAIL in both directions, and the repairs stopped

Ground truth taken from the fool's own prompt text (written by `tarbuckle-*.py`, not by the function
under test): **24 known mumbles, 41 known non-mumbles**, of 65.

- **must-detect 22/24** — two mumbles return `human_turns == 1` and read as human-attended.
- **must-not-flag 35/41** — six non-mumbles return `0`, and ⚠ **at least one of those is CORRECT**:
  `b7e7eb39` is the unattended executor session of 2026-08-31 (PENDING-172), which genuinely has no
  human turn. **`0` never meant "mumble". It means "nobody spoke"**, which is equally true of an
  unattended real session.

**Mechanism, confirmed rather than inferred.** `human_turns()` skips any `user` record containing one
of seven `NONHUMAN` markers. A mumble embeds the *previous session's* material in its prompt, so it
inherits those markers **only when the session it was mumbling about happened to contain a slash
command**. Of 24 mumbles, 22 embed a marker and are excluded; **2 embed none — and those 2 are
precisely the 2 that leak.** The correlation is content-dependent coincidence. The function's own
docstring says it answers *"did an executor just run unattended?"*, and at that job it is correct.

**No replacement was built, per the handover and for its stated reason:** three controls passed and a
fourth broke, so the finding is about the **control set**, not only the function. All three test the
question the function was *for*; none could see the question it was being *reused* for.

### GATE 2b — composition stands; two terms do not close

Reconstructed from the archive by the same ground-truth signature:

| date | N | real | mumble | % |
|---|---|---|---|---|
| 2026-08-31 | 52 | 41 | 11 | 21.2 |
| 2026-09-01 | 50 | 39 | 11 | 22.0 |
| 2026-09-03 (measured live) | 65 | 41 | 24 | 36.9 |

**-178's composition claim stands**: 11 mumbles and "roughly a fifth" close exactly on 08-31. Against
-178's enumerated `54 = 43 + 11` on 09-01 I reconstruct `50 = 39 + 11` — **the mumble term identical,
the whole 4-file gap in the real-session term.** ⚠ **Not asserted against -178.** My method models the
prune as a 30-day window over `source_mtime` and is a demonstrated **lower bound**: preservation has
run **twice only** (08-26, 09-03). Where we disagree, -178 enumerated live and is better positioned.

**Record-only correction applied** (authorized): `MEMORY.md` carried *"N-now 44/84 as of 2026-08-31,
DOWN 7 from 51"* — wrong in the number and **backwards in the direction**. Now **65/84 measured**,
with composition stated, and the stale `governance-drift-check.py:331` pointer corrected to `:513`
(`:331` is a register-check control; both lines read before changing it).

### GATE 2c — five sites in four files; -178's scope holds

Controls named before the run. Enumerating consumers: `governance-drift-check.py` (426, 513) ·
`wake-digest.py` (354 + the selftest sample) · `tarbuckle-invoke.py` (37, 44) ·
`preserve-transcripts.py` (45, 126, 172). `tarbuckle-wrap.py` and `-seam.py` are **not** consumers —
they receive `transcript_path` from the hook payload.

⚠ **What the census cannot see, as output rather than caveat:** (i) a consumer that builds the path by
component join *and* never enumerates `*.jsonl` on a matching line — **demonstrated: `wake-digest.py`
does exactly this and escaped a literal-fragment grep over the whole fleet**; (ii) anything reaching
the population through a hook-supplied `transcript_path`; (iii) anything outside the swept roots.
**(i) applies retroactively to the four-site census inherited from the previous wrap, which was
grep-derived by the method just shown to be blind.**

### Filed, corrected, committed

- **PENDING-179** — as an item, not a `-178 ADDENDUM 1`, on PENDING-145's mechanism: `ruled_pendings`
  claims a **number**, so an addendum would be suppressed the moment -178 is ruled. The undecided
  filing moratorium is disclosed inside the item, with the previous session's contrary reasoning
  preserved rather than overridden.
- **`c150bdf`**, pushed to `github` and `gitea`, **both verified at the same SHA**.
- **OWED-5 queued** in PENDING-141's owed-entries list (the ratified queue, REVIEWED-123 cond. 3).

### ⚠ The archive is not backed up, and the script says it is

Found while checking repo state for the commit. `preserve-transcripts.py:11` states it copies
transcripts *"to a git-tracked location so the population stops shrinking."* **`~/_Dev/claude-transcript-archive`
has no `.git`, no parent repo, and is 144 MB.** The copying works — every file is re-hashed on
readback — but the sentence describing where they went is false, and **11 transcripts now exist in
exactly one place on one disk.** PENDING-144's class (substrate claims inside governance scripts
checked by nothing), at the worst possible site: the docstring is what anyone reads to decide whether
the evidence is safe.

## PRESENT — how it stands

**The mood.** Unusually clean, and the cleanliness is entirely borrowed. The handover did the work
that made this session good: it put preservation before the control, excluded site 1 so I could not
wander into it, and — decisively — insisted the "already controlled" claim be tested before anything
was wired anywhere. Left to my own plan I would have written a positive control for a fixture and
wired a broken predicate into three more sites, all of it green.

**What was corrected — three, and the first is mine from four hours earlier.**
1. **My wake census used `<= 1` as the mumble threshold**, silently absorbing 3 transcripts that
   return exactly 1. The real distribution is 28 at 0 and 3 at 1, and the ground-truth split is
   24/41. The wake number (31) was wrong and the method behind it was wrong in a more interesting way.
2. **A vacuous `PASS (0/0)`.** My first ground-truth query used the signature from
   `tarbuckle-invoke.py` — the file I had just read — and matched **0 of 65**. The gate went green on
   an empty denominator. Caught by *a null search is evidence about the QUERY*; opening a transcript
   showed the mumbles come from three **other** fool surfaces (`tarbuckle-mumble/-wrap/-seam`) that
   the inherited site census never named. **This is now OWED-5.**
3. **The previous wrap predicted the digest selftest would "fail on every run from now on."**
   Falsified within 2.2 h: it passed at this wake (5 of 13). The defect is *intermittent* — a
   function of the mumble rate in a rolling 13-transcript window — which is harder to notice than a
   permanent failure.

**Confidence to recalibrate.**
- **Verified by running it:** the 2a confusion matrix over all 65 with independent ground truth; the
  leak mechanism (2 leaks == 2 marker-free mumbles, exactly); N-now 65 and its 24/41 composition;
  preservation readback; both remotes at `c150bdf`; the archive has no `.git`.
- **Reconstructed, lower-bound, NOT asserted against -178:** the 08-31 and 09-01 rows of the 2b table.
- **Inherited and now known to be unreliable:** the four-site census from the previous wrap —
  grep-derived by a method this session demonstrated is blind to component-joined paths.

**Instruments:** 4 run (ground-truth classifier · archive reconstruction · consumer census ·
histogram) · **1 carrying a control written before first execution** (the consumer census, whose
must-find and must-not-find were named in the file before it ran — and the must-not-find *fired*,
catching the vendored-docs pollution) · **K = 0** — none duplicated anything banked; all four answered
questions asked once. ⚠ **The census was narrowed three times.** Repeated narrowing can end by
confirming the sites it was built around; the negative control and the explicit blind-spot statement
are what keep that honest, and they are not a substitute for someone checking it.

**Decisions deferred, and why.**
- **No repair at any site**, and no replacement discriminator — gate 2a's stated consequence.
- **`governance-drift-check.py:513` not examined at all** — excluded on receipt.
- **Did not `git init` the archive.** 144 MB is real weight for both remotes, and full transcripts
  contain everything ever typed in these sessions; whether they belong on a hosted remote is a
  steward decision, not a chore.
- **Did not correct `preserve-transcripts.py`'s false docstring.** Editing a governance script's
  stated rationale to match a worse reality is the direction that makes docstrings worthless. It
  should become true, or be corrected as a disclosed change.
- **OWED-5 queued rather than patched in.** Turning the rule on converts currently-green controls to
  red across the fleet; that is a ruling, not an edit — and the ladder is frozen regardless.

## FUTURE — what pulls

**The pulling thread is unchanged in name and now has two steward-set acts in front of it.**
The fr cell's last two steps still pull. But the steward has set the next session's opening
explicitly, and it is not that.

### Next session, first work — steward-directed 2026-09-04

1. **`git init` the transcript archive.** The steward has named this as the wake's first act.
2. **Suspend the automatic grading at `trigger_fired()`.** ⚠ **Stated precisely by the steward, and
   the precision is the point: this is NOT a repair, NOT a filter, and NOT -178 option (b). It
   changes nothing about what the counter counts, so the site-1 exclusion survives intact.** It
   prevents a grade firing on a population whose contamination is measured and rising.

### Then the joint ruling — steward, 2026-09-04

**-178 and -179 are ONE decidable unit** and will be ruled together. -179 is -178's evidence and says
so in prose, because the numbering cannot carry it (PENDING-145). The joint ruling covers **unit,
window, recurrence, predicate, and seeding.**

⚠ **Seeding has grown a second half, and it is new today.** Preservation has **permanently diverged
the two stores**: 11 files exist only in the archive. So the question is no longer only *what N*, but
**over which store** — and the trigger reads **live**, while any honest grading of a post-08-07
population must read **preserved**.

**Steward's positions, recorded so they are not re-litigated:**
- **Agrees with -179's option (c)** — a mumble that declares itself cannot be misread by a marker
  coincidence — **and with the refusal to build it now.**
- **The lesson is the control set, and it generalizes past this function:** three controls passed and
  a fourth broke, and all three tested the question the function was *for* rather than the question it
  was *reused* for.

**Other horizons, ranked.**
- **The archive's single copy** — a local second copy is the cheap half and answers none of the
  remote question. Unresolved.
- **`preserve-transcripts.py:11`'s false claim** — becomes true after the `git init`, which is the
  clean resolution and is why the steward put the init first.
- **`~/.claude/agents` under version control** — steward-directed, still owed, before anything else
  edits it.
- **The filing moratorium** — undecided, and PENDING-179 was filed under it with that disclosed.
- **PENDING-171** → unblocks 322 unread of 362 · **-160** · **-168** · the §5 regrade, an eighth session.
- **Parked worker `acaabadf`** — stopped, still carrying `--reply-on-resume`. Untouched again.

**Pause statement.** I am about to be away from this and the context is being cleared deliberately.
What I want to find still pulling is the fr cell's last two steps — but what I want the next session
to *do first* is the two acts above, in that order, because the steward set them and because the
`git init` is what makes an already-false docstring true rather than requiring it to be edited into
honesty. ⚠ What I want the next session to **notice** is that this session's entire value came from a
gate that stood between a plan and its execution. The plan was mine, it was confident, and it was
wrong. **The handover did not improve the fix; it prevented it.**

**Literal question for next-Claude** *(checkable; turns on the record, not introspection)*:
**How many other must-detect controls in the fleet currently have a denominator of zero?** OWED-5 is
queued on a single observed instance. The fleet's controls are enumerable and their denominators are
computable without changing any of them. ⚠ If the answer is zero, OWED-5 is a rule earned from one
accident and should be labelled that way rather than carried as a general finding. If it is more than
zero, then some number of currently-green controls have never tested anything — and nobody knows
which, because a vacuous pass and a real pass print the same word.

---

## ADDENDUM 1 — after the wrap: the jurist corrects the record, and one claim in it was wrong

⚠ **Appended 2026-09-04 after the session record was written, committed and pushed.** The jurist read
the wrap and returned five corrections, two operational. Filed into `PENDING-179 AMENDMENT 1`; the
substance is repeated here because this file is what the next wake reads.

### The correction that matters most — I gave credit to the wrong thing

The wrap said *"the handover's ordering, at every point it was load-bearing."* **The gate's POSITION
held; its CONTENT did not.** Gate 2a as specified was *one arm, one transcript, expected 0*. **Run as
written it greens** — 22 of 24 mumbles return 0, so a single sampled mumble passes with ~92%
probability and the repair proceeds. The finding exists because I **replaced the specification**:
ground truth taken independently of the function under test, across the whole population, plus a
`must-not-flag` arm nobody asked for.

⚠ **"Handovers are load-bearing, follow them" is the wrong lesson and the more comfortable one.**
*A gate's existence bought the chance to catch this; improving on the gate's specification is what
caught it.* A session inheriting the comfortable version runs the next gate as written.

### A NEW measurement that supersedes this session's own origin claim

The 2026-09-03 record said the failing selftest sampled 13 transcripts and *"all thirteen are
Tarbuckle mumbles"*, with mumble-displacement as the cause. **Measured live on the actual slice:**

- **1 real session, 12 mumbles** — not 13 mumbles.
- **verdict `wrapped`: 1 from the real session, 4 FROM MUMBLES.**
- the predicate is `"wrapped" in _v` — **satisfiable by mumbles alone.** It would pass with **zero**
  real sessions in the slice.

⚠ **Its label claims "a real session reads as WRAPPED end-to-end"; its test asks whether any
transcript whatever drew that verdict.** That is **OWED-1's wrong-subject family, inside the control
that was supposed to be evidence about mumbles.**

### And the question I still have not answered

The jurist asked for **the date of the control's first failing run**, because a start predating
2026-08-25 would mean the mumble is not its only cause. A reconstruction over the preserved archive
found no date from 08-20 to 09-04 on which the slice held zero real sessions — **but that
reconstruction is not sound for the purpose**: manifest `source_mtime` is snapshot-time rather than
current, and the verdict also depends on `wrap_events()` git history that was not replayed.
⚠ **OPEN, not answered.** The wrap's *"intermittent"* was a better-founded description **substituted
for the question** — the same move twice, and naming it is the correction.

### Two operational changes for the next session

1. **UNBUNDLE the two acts, suspension first.** The wrap's own *decisions deferred* records the
   `git init` as a **steward call**, so it may wait days — while the suspension is the act with a
   **firing distance** (N=65/84, 19 away, 13 net in three days). Different files, no dependency;
   serializing puts the deadline behind the deliberation.
2. **THE SUSPENSION MUST CARRY A REPLACEMENT BOUND IN THE SAME ACT.** REVIEWED-123 cond. 2 installed
   *grading at 84* as the freeze's bound — *"a hold with no expiry and no visible distance to expiry
   is how a temporary freeze becomes a permanent one."* Suspending grading **removes exactly that
   bound.** Tie the replacement to the joint ruling or to a working session predicate; the 09-16
   report obligation survives either way. ⚠ Bites **OWED-5**, now long-coupled to a ladder trial.

### Also carried, and also mine to have missed

**Gate 2c's census limitation is in the wrap but was not in PENDING-179** — which is what gets ruled
on. It was narrowed three times, and repeated narrowing can terminate by confirming the sites the
predicate was built around. Now in AMENDMENT 1; **read 2c's "no additional consumers" as the weakest
of the three findings.**

**The moratorium is still undecided at 64 open items**, and it is shaping filing in **both**
directions — one finding withheld on 09-03, one filed under protest on 09-03. That is the worst
state for an undecided rule to sit in.

### What the jurist credited, recorded because credits are evidence too

The vacuous `0/0` was caught by a **standing rule** (*a null search is evidence about the query*),
not by luck — the class was instrumented. Declining to edit the false docstring, with the reason
given. And OWED-5's posture: a rule resting on one observed instance, carrying a stated test that
could retire it as an accident.

---

## ADDENDUM 2 — the `[FIX]` warrant is void, the date question is retired, and OWED-5 is narrower than it looked

⚠ **Second jurist pass, same day.** Three corrections; the first changes what the joint ruling must cover. Filed as `PENDING-179 AMENDMENT 2`.

### The `[FIX]` warrant is void at every site — traced, not accepted

The 2026-09-03 reclassification rested on one argument: *the label says "a real session reads as
WRAPPED end-to-end", the sample contains no real sessions, so restoring the population repairs the
instrument against its existing specification.* Read at source:

```
_v = [wrap_verdict(transcript_span(p), _ev)[0] for p in _tx[-14:-1]]
chk(f"a real session reads as WRAPPED end-to-end  [{_v.count('wrapped')} of {len(_v)}]",
    "wrapped" in _v)
```

**`"wrapped" in _v` does not restrict to real sessions anywhere.** Restoring the population repairs
nothing: the test passes on mumbles alone and would pass on a correctly-populated slice in which
**every real session failed**. ⚠ **The defect was never the sample.** Label and test disagree about
*subject*; nothing that changes what enters the slice reconciles them.

Sites 2 and 4 never had a quoted specification. Site 3's is contradicted by its own implementation.
Site 1 is excluded. **Nothing in the census is `[FIX]`-warranted**, and whatever replaces the
selftest is a **rewrite of the predicate against its label** — larger than the act reclassified.

⚠ **Withdrawn explicitly in three places** (this record, `MEMORY-reference.md`'s archived block by
annotation-not-edit, and the 09-03 session record) because *"sites 2–4 are `[FIX]`, merely gated"* is
exactly the premise a later session inherits without re-deriving.

### The date question is RETIRED BY FINDING — AMENDMENT 1 recorded it open; that is withdrawn

It was wanted to test one inference: *a start predating 2026-08-25 would mean the mumble is not the
only cause.* The predicate finding settles it **without the date** — a test asking whether *any*
transcript wrapped has never tested its label **since it was written**, necessarily predating
mumbles. **The mumble is not its cause at all**, only a change in what fills a slice whose
composition the test was already ignoring. ⚠ **Do not replay `wrap_events()` git history for it.**
What the intermittency tracks is total wrap-verdict density — evidence about nothing in -178 or -179.

### OWED-5 sees one of two vacuity classes

The selftest had **thirteen files in its denominator**. It passes the denominator rule and tests
nothing about its subject regardless.

| class | shape | caught by |
|---|---|---|
| empty-set vacuity | no cases at all | **OWED-5** |
| wrong-subject vacuity | full denominator, predicate ≠ label's subject | **OWED-1 — invisible to OWED-5** |

⚠ The fleet denominator sweep has a **harder and more common sibling**: *how many controls have a
predicate that does not implement the subject named in their own label.* The selftest is the
demonstrated instance, making this a **recurrence of OWED-1, not a discovery**.

### On the correction pattern itself — worth banking for PENDING-89

**Three of PENDING-179's four load-bearing claims were corrected within 24 hours of filing**, by the
jurist or by re-measurement. That is the configuration working. It is also a reason to read the
remaining claims as provisional.

⚠ **And one correlated miss, which is PENDING-89's actual question.** The jurist declined credit for
self-correcting on urgency, noting the correction only ran because a measured N of 65 arrived, and
that the error was a *class* error — reading the absence of a stated rate as the absence of urgency.
**The executor shared that miss.** Both items correctly withheld a rate; both parties read the
withholding as slack; neither flagged the firing distance until a number was measured. **What one
missed, the other missed too, in the same direction, for the same reason.** Recorded per
Constitutional Constraint 6, which requires evidence against the differently-biased-checkers doctrine
to be logged when observed rather than only when sought.