[HARDENING] Close PENDING-82 and -118; mark 119/120/123 BUILT

82 is discharged by events. Its 2026-07-28 substrate check said there was no
mcpServers key; today the config carries mcpServers: governance, and the jurist
used the tools in three consecutive rulings — opening graduation-spec directly
and refusing to rule from my summary, which is the capability the item existed
to create. Two residuals carried, not buried: the read enum reaches neither the
runbook nor the R0 contract, and the installed surface has 8 keys and a search
tool the description does not name.

118 is built, and building it REFUTED the option I had recommended. I wrote that
the checker already parses the archive format. It does not — the marker is an
HTML comment and there are zero in either register file; their deferrals are
prose, 53 and 26. Widening alone would have scanned two more files, found
nothing and reported clean: a silent net built to close a blind spot, which is
the failure the item was filed to describe.

So the widening ships with its limit in its own output — prose deferrals counted
and reported un-machine-readable, never as absent, with counting explicitly not
classifying. The census stays owed.

119/120/123 marked BUILT with their commits so the built-vs-ruled checker sees
them; all three were already ruled, so this closes a reporting gap, not an
authorization one.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A35wiD55yRHj5U1ECZAX4t

PENDING.md

@@ -496,7 +496,19 @@ Merge that `mcpServers` key into `~/Library/Application Support/Claude/claude_de
 Delete the Cowork party entry and every reference to `COWORK.md`. Grounds, now two: a third executor costs a third doctrine copy of a document that is `CLAUDE.md` with the nouns changed; and Cowork could not have served as the jurist's filesystem eyes even in principle, since `coworkUserFilesPath` points at `~/Claude`, which does not exist, and remote Cowork — the incoming default — runs no local MCP at all.
 
 **Files affected:** new `~/dotfiles/scripts/governance-mcp.py`; `~/dotfiles/scripts/wake-digest.py` (`item_spans()` fence-awareness + 3 controls). Awaiting steward hand: `claude_desktop_config.json`, Claude.app §Your Role and §Standing Context.
-**Awaiting:** Steward authorization to install the `mcpServers` key. The server itself is inert until then — nothing loads it.
+**Awaiting:** ~~Steward authorization to install~~ → **INSTALLED AND IN USE. CLOSED 2026-08-08.**
+
+---
+
+### AMENDMENT 1 — 2026-08-08, discharged by events; the item's substrate check is stale
+
+**§A — INSTALLED.** The item's 2026-07-28 substrate check recorded *"`claude_desktop_config.json` has **no `mcpServers` key**"*. Verified today: top-level keys are **`mcpServers`, `coworkUserFilesPath`, `preferences`**, and `mcpServers` contains **`governance`**. The gate this item waited on has been passed.
+
+**§B — And it is not merely installed, it is LOAD-BEARING.** Empirically, in a single day: the jurist used `governance_read`, `governance_item`, `governance_search` and `governance_state` across **three consecutive rulings**, opened `graduation-spec.yaml` L1–60 directly, and **refused to rule from the executor's summary** — *"ruling from the executor's summary of its own mandate is exactly the shape I should refuse."* That refusal is the capability this item existed to create, exercised. It also caught, from substrate the executor had quoted, an adverse ratified ruling the executor had missed (REVIEWED-53).
+
+**§C — ⚠ Two residuals, carried not buried.** (1) **The `governance_read` enum does not reach everything a ruling may rest on** — `conversion-runbook.yaml` and the R0 contract are reachable by **no key**, and one ruling had a leg in executor testimony until the steward relayed the files by hand. The remedy is its own extension mechanism: **the server takes keys from a list, so extending the list is the fix.** Recorded in this item's earlier amendment; it survives this closure as a **named follow-on**, not a reason to hold the item open. (2) The installed surface differs from the description above — **8 keys, not "six enumerated documents"**, and a `governance_search` tool the five-tool list does not name. The description is stale; the substrate is authoritative.
+
+**Closed:** the proposal was to build and install a read-only substrate interface for the jurist. It is built, installed, used, and has demonstrably changed rulings. What remains is a **bounded extension of an existing, working thing**, which is a different item.
 
 ### AMENDMENT — 2026-08-08, a concrete enum, from a ruling that hit the wall
 
@@ -1683,7 +1695,30 @@ This is harmless today only because voice ⟺ source: measured, **max distinct v
 
 **Files affected:** `~/dotfiles/scripts/governance-drift-check.py`; a one-time census artifact (home to be decided with the ruling).
 
-**Awaiting:** Steward authorization.
+**Awaiting:** ~~Steward authorization.~~ → **BUILT 2026-08-08, `see dotfiles HEAD`. ⚠ AND THE ITEM'S OWN OPTION (1) IS REFUTED BY BUILDING IT.**
+
+---
+
+### AMENDMENT 1 — 2026-08-08, built — and option (1) rested on a false premise about the format
+
+**§A — ⚠ MY OPTION (1) WAS WRONG, and building it is what showed that.** I wrote: *"Widen the scan to `~/PENDING-archive.md`. Smallest change; **the checker already parses that exact format.**"* **It does not.** The structured marker is an HTML comment — `<!-- DEFERRED-DECISION: slug … -->` — and there are **ZERO** of those in `PENDING.md` **or** in `PENDING-archive.md`. Measured 2026-08-08. Their deferrals are **prose**: **53** occurrences of `defer*` in `PENDING.md`, **26** in the archive.
+
+**⇒ Widening alone would have scanned two more files, found nothing, and reported clean** — *a silent net, built to close a blind spot.* That is precisely the failure class this item was filed to describe, and I had specified it as the remedy.
+
+**§B — So the widening ships WITH its own limit stated in the output.** Structured blocks are now found anywhere in the register; prose deferrals are **counted and reported as un-machine-readable, never as absent**:
+
+```
+✓ deferred decisions: 2 tracked, none due (2 checkable, 0 manual-only)
+  ⚠ plus 79 PROSE deferral mention(s) in the register (PENDING.md 53, PENDING-archive.md 26) — these carry no
+    DEFERRED-DECISION block, so NO trigger is machine-checkable for any of them.
+    Counted, not classified. Whether any condition has fired is unestablished.
+```
+
+⚠ **Counting is not classifying.** 79 is an upper bound on candidates, not a count of deferrals — the regex matches any use of the word. **How many carry a condition, and how many of those have fired, is a READING task** and is reported as unestablished rather than skipped. That is the honest version of what option (2)'s census asked for, and the census itself remains **owed**.
+
+**§C — Three controls added**, per the script's standard: the prose counter fires on a known-present phrase, stays silent on unrelated text, and the register files are provably inside the widened scan.
+
+**§D — What this closes, and what it does not.** **Closes:** the checker no longer reads only `docs/**` — a structured deferral filed anywhere in the register is now seen, and the register's prose deferrals are **visible as a named unknown** instead of invisible. **Does not close:** the classification. The item's own ⚠ said *"size unmeasured, deliberately"*; it is now **bounded and still unclassified**, which is a better state and not the finished one.
 
 ---
 
@@ -1730,7 +1765,7 @@ This is harmless today only because voice ⟺ source: measured, **max distinct v
 
 **§D — CONDITION ON (i): declare the cost threshold now, with its action.** `0.218 s over 14 sources` is honest about being burst-sized; (e) is unconditional and scales with sources × file size. **When it exceeds ~1 s, (e) re-scopes or hands off to (a)'s scheduled job.** Stated now because *a per-commit cost that grows unremarked converts a tripwire into a `--no-verify` habit* — this thread's own failure class arriving by the back door.
 
-**Awaiting:** placement of the ruling. Build on placement: one line in `.precommit-triggers`, the §D threshold recorded beside it, tagged REVIEWED-N.
+**Awaiting:** ~~placement of the ruling~~ → **BUILT 2026-08-08, `2534dfb`** under REVIEWED-102. One line in `.precommit-triggers` (`. | python3 engine/ingest_gate.py --check-only`), unconditional, with the §D cost threshold recorded beside it. Acceptance: both rules fire in declared order, cheapest first.
 
 ---
 
@@ -1782,7 +1817,7 @@ So the pathspec was **not silence — it was a cost commitment inside the author
 
 **§D — Interaction with PENDING-119, if both land.** `.precommit-triggers` would carry two lines with overlapping paths; an engine commit pays ~2 s (fleet) + 0.218 s (binding). **Declare the order in the file** so a red is attributable to one check without reading both.
 
-**Awaiting:** placement of the ruling.
+**Awaiting:** ~~placement of the ruling~~ → **BUILT 2026-08-08, `2534dfb`** under REVIEWED-103. Pathspec widened to `corpus/ engine/ tests/ scripts/run-fleet.sh`. Acceptance decomposed per condition 2: `eecc8bb` replayed (both files match); red direction refuses — **fixture SYNTHETIC and labelled**, no real red `engine/` commit exists in 24 candidates; docs-only runs no suite.
 
 ---
 
@@ -2043,7 +2078,7 @@ Three cases, all discriminated: a rule ran → existing output already says so,
 
 **§E — Related doctrine, filed as PENDING-124.** This item's *"needs a third state, not a pass or a fail"* and PENDING-122's `cannot-assess` are one finding reached twice in one day.
 
-**Awaiting:** placement of the ruling. **Build order on placement: 123 → 119(i) → 120(a).**
+**Awaiting:** ~~placement of the ruling~~ → **BUILT 2026-08-08, `448ce37`** under REVIEWED-105, first in the ruled order. (b)+(e): malformed declarations refuse with file/line/fault/`--no-verify`; a triggers file declaring nothing reports itself unguarded; the per-rule line prints in exactly the ambiguous case. Matched-rule output byte-identical. All seven table rows non-silent.
 
 ---
 

REVIEWED.md

@@ -1293,4 +1293,262 @@ blocks without saying why is replaced by habit within a week.
 4. SEQUENCING across the four open items: land 123 before 119(i) and 120(a). Both add lines to .precommit-triggers; a validator that catches a
 malformed line should exist before the file grows. In the other order, the first thing to test the new declarations is the declarations themselves.
 
-**If AUTHORIZED:** Proceed with (b) + (e). Build order: REVIEWED-105 → REVIEWED-102 (i) → REVIEWED-103 (a). Tag commits with REVIEWED-105.
+**If AUTHORIZED:** Proceed with (b) + (e). Build order: REVIEWED-105 → REVIEWED-102 (i) → REVIEWED-103 (a). Tag commits with REVIEWED-105.
+
+## REVIEWED-106 — PENDING-124 — A check whose subject can be absent cannot be two-valued
+**Date:** 2026-08-08
+**Decision:** DESIGN GATE PASSED WITH CONDITIONS. Q1 applied, not extended — no constitutional change, no [ESCALATE]. Landing: one ladder entry.
+Five conditions, none changing the doctrine's substance; three changing what is recorded about its basis.
+**Stores:** claude-md L251-257, REVIEWED-104, REVIEWED-105, PENDING-124 verbatim via the governance tools. R0 §3 verified directly against the
+sha-matched contract (db39a503…, L180), so I.4 is substrate for this jurist, not testimony — the package's front-matter was true two turns ago and
+is now false.
+
+1. Q1 APPLIED. Constraint 4 has two clauses. The contrary reading engages only the second; the first speaks of limits, not failures, and "I could
+not look" is a limit. The executor overstated its own uncertainty here, and deserves to know the replacement is firm.
+
+2. CONDITION 1 — the quote-verification pass gains its own third state before it is relied on again. It reported verified on a normalized
+reconstruction of REVIEWED-104, which had also dropped the sentence answering the package's own Q2. A two-valued verifier inside a package arguing
+for three-valued verifiers. This is also a tenth instance, and the only one in the set independent of the advisory that proposed the doctrine —
+produced by the gate, not the proposal.
+
+3. CONDITION 2 — the ladder entry names the attested-absence family and cites the 2026-07-05 ruling (REVIEWED-47), so it joins a lineage rather
+than standing as an orphan under a constitutional clause.
+
+4. CONDITION 3 — the evidence statement is corrected before placement: two pre-existing instances of the shape, three of the adjacent principle,
+one found at the gate. The uncorrected "five … same shape" would have been inherited as load-bearing.
+
+5. CONDITION 4 — III.1 carries the environment-vs-defect split in the normative statement, not only in the consequence-trace. CONDITION 5 — landing
+as leaned: one named instrument on reference-verification-ladder.md, nothing in ~/CLAUDE.md.
+
+6. Q2 binds at reporting AND aggregation; the aggregation half is already ruled by REVIEWED-104 §1's closing sentence. Q3 names left free, with the
+ratified test recorded: one name per referent, not one per concept. Q4 ratify not provisional, and not on the count — four of nine instances are
+downstream of the advisory that proposed the doctrine; the chamber census is owed, not blocking.
+
+**If AUTHORIZED:** Conditions 1-5 are discharged as of 2026-08-08. Tag with REVIEWED-106.
+
+## REVIEWED-107 — PENDING-125 — A live false attestation in the governed record: Mauss's reading_index_status
+**Date:** 2026-08-08
+**Decision:** AUTHORIZED — option (a). D-1 lane, authorized verbally; RECORDED RETROSPECTIVELY, after the build.
+**Notes:** (a) and (b) are different acts and were correctly separated. Correcting the field stops the record asserting something false today and
+needs no ruling; re-anchoring the index is curatorial work that must carry the ladder's re-anchor = re-verify discipline, since re-anchoring
+without re-verifying is what produced this class. (b) remains open.
+
+1. Built as 8231bce. reading_index_status VERIFIED-BOUND → SHA-STALE, bounded to one field, shas untouched, manifest re-parses at 14 sources.
+
+2. The executor checked the sub-question it had flagged before choosing a value, and the answer changed the entry: the vocabulary is UNDEFINED.
+Three tokens in use across the manifest, no definition anywhere in either repo, every external mention prose about this defect rather than a
+specification. SHA-STALE is therefore a fourth undefined token, added because none of the three could state the truth, and recorded as a known cost
+rather than minted quietly.
+
+3. Noted: the commit was the trigger mechanism's first real corpus exercise — both declared rules fired, fleet green, not a probe.
+
+**If AUTHORIZED:** Tag with REVIEWED-107. (b) stays open under PENDING-125.
+
+## REVIEWED-108 — PENDING-126 — Two holes in the fleet, and the census that followed
+**Date:** 2026-08-08
+**Decision:** AUTHORIZED — options (a) then (c). D-1 lane, authorized verbally; RECORDED RETROSPECTIVELY, after the build.
+**Notes:** Merged with PENDING-122 on the executor's argument that they are one subject with overlapping files, and that hole 2 was a prerequisite
+for verifying 122 — while suites raise, cannot-assess cannot be told from red. The merge was right.
+
+1. (a) built as 8ff5a9f. Hole 1: close_ranges' section_end bound guarded, both branches; discriminating negative run — bound removed → two named
+failures citing 499 and 400, restored → 47/47. Hole 2 was THREE sites, not the one filed; fixed as a class. An induced citability break went from a
+single StopIteration traceback to seven named failures.
+
+2. (c) the census, done, and it renamed the class. Crash-rather-than-name: 3 of 7 suites under 3 triggers, origin suite-side direct access. The
+unguarded-rule question is unanswerable by inspection — token-mention said 13 of 13 touched, which is worthless since hole 1 lived in a touched
+clause. Mutation: 4 of 7 caught, and all 3 survivors verified EQUIVALENT on current data by sentinel and by positive control. So hole 1 was never
+an unguarded rule; it was a guard the live corpus cannot exercise, and three more of that shape exist in R0 alone. Latent, not wrong.
+
+3. The crash class then closed at the preflight rather than at six sites (d21a43b), on the ground that all six depend on one invariant. Closing the
+six revealed two more under a fourth degraded state, in suites the census had cleared: 6 → 8. The eighth originates in ENGINE code (retrieve.py
+_work_map) and was deliberately not fixed by a test-side patch; whether retrieve() should degrade rather than raise is filed as an engine question
+rather than hidden.
+
+4. Final census, five degraded states: zero crashes, against 3/3/0/3 before.
+
+5. ⚠ Recorded as owed, not blocking: the census covered R0's clauses only. n0/n1/v0/v1/cluster-a are unmutated and their escape rate is unknown,
+not zero.
+
+**If AUTHORIZED:** Tag with REVIEWED-108.
+
+## REVIEWED-109 — PENDING-127 — R0 §4 emits two states where §3 rules three
+**Date:** 2026-08-08
+**Decision:** AUTHORIZED — option (a). D-1 lane, authorized verbally; RECORDED RETROSPECTIVELY, after the build.
+**Notes:** The item said §4 was binary against §3's three states. It was worse: the CODE was unary. emit promoted baseline_sha256 to content_sha256
+on every region — 261 of 261 for Alexander, including regions no instrument had verified — under a hardcoded date. Three different answers lived
+in one contract and one module.
+
+1. Built as ccc4d6c. Contract v0.1 → v0.2, superseded sentence preserved in place. Suite 31 → 44. Fleet 7/7.
+
+2. The fix goes further than (a) asked, on the item's own logic, and that is ratified: a content_sha256 attests the whole span while name-landing
+is evidence about the anchor's first line, so recording the former because the latter held promotes a weaker claim into a stronger one — the
+PENDING-47 shape. Emission now records NO new fingerprints; one enters only through an attested re-verification. This is what makes PENDING-121
+condition 4 reachable rather than aspirational.
+
+3. State determination is now one function called by validate and emit, which had silently disagreed — §3's own "one mechanism with two call sites"
+applied to the module's interior.
+
+4. ⚠ The acceptance fixture named in the item was STALE, and measuring corrected it. Alexander's five front_matter anchors were partitioned out on
+2026-08-07; Alexander is 261/261 name-landing with zero unverified. The real unverified population is Mauss 23 and after-the-reply 33. The
+discriminating pair is therefore Alexander against Mauss — two real artifacts, a better control than the one specified. stale is unreachable from
+live data and its control is labelled synthetic.
+
+5. One pre-existing check went red and was replaced rather than deleted: it asserted the promotion this item rules a defect. A test that pinned the
+old contract is evidence of what the contract used to say.
+
+**If AUTHORIZED:** Tag with REVIEWED-109.
+
+## REVIEWED-110 — PENDING-121 — engine_source_binding: prose to declared surfaces, and the fingerprint specified but never recorded
+**Date:** 2026-08-08 (design gate 2026-08-08; two follow-on passes same day) · placed retrospectively
+**Decision:** DESIGN GATE PASSED WITH CONDITIONS (1-4), then HELD OPEN for a redraft of Part IV.2 after substrate
+verification. The redraft is filed (item Amendment 3 §D) and is NOT gated by this entry. Placement gate outstanding,
+jointly with PENDING-128.
+**Stores:** REVIEWED-101, REVIEWED-53, REVIEWED-47, PENDING-121, PENDING-119, PENDING-120, PENDING-47 verbatim via the
+governance tools; graduation-spec.yaml L1-60 via governance_read. conversion-runbook.yaml and the R0 contract were
+unreachable at the time of ruling and were read only after the steward supplied them, sha-matched against the request's
+table (fffeed86…, db39a503…). engine/reading_index.py was never read by the jurist: condition 4 was ruled
+prophylactically and REVIEWED-109 later found the code unary, which is worse than the item claimed.
+governance_state() timed out; no drift count or repo status informed this ruling.
+
+1. AN ADVERSE RATIFIED RULING WAS MISSING FROM THE GROUNDING PASS. REVIEWED-53 kept engine_source_binding as ONE entry —
+'names a relationship across three files that move together; fragmenting recreates the failure' — and the package
+proposed five siblings. REVIEWED-101 condition 1 did not cite it either: two rulings from one lane pointing opposite
+ways, neither aware of the other. Treated as a verification trigger, not a precedence call. Resolution: the entry stays
+ONE, surfaces become addressable members, and the co-movement invariant becomes declared data rather than why: prose —
+because prose is what this amendment establishes no consumer reads.
+
+2. CONDITIONS IN FORCE. (1) co-movement declared, a consumer verifying a proper subset reports incomplete never clean —
+and this collapses with IV.1 para 2 into one requirement, drafted once. (2) resolve scope, then DERIVE the enumeration
+from the runbook's list plus the per-region surface, justifying every omission; never compose afresh. (3) no dated counts
+in declared data — locators and semantics only, population computed at read time; unverified-by-construction survives
+only as a rule, since a rule does not go stale and a count does. (4) the promotion rule is REVIEWED-47 applied, not new
+normative text: reuse the ratified by/against/result attestation shape under the single shared guard, reducing the
+constitutional change to one requirement.
+
+3. Q1 — the zero-evidence surface is CARRIED, never omitted, and the stronger form is adopted. The marking is not what
+prevents the Part V inversion; not computing a heterogeneous aggregate is. Report per surface; any derived aggregate
+reports unverified when a member is unpopulated or unchecked. Two discriminations required: the drift signal must be
+separable from the unverified signal, or the check becomes a constant nobody reads; and '271 verified' must never
+surface unqualified, since R0 §3 holds name-landing insufficient.
+
+4. Q2 — reading_index_status is NOT a binding surface and NOT evidence, and no consumer of this enumeration may read it
+as such. Retention, demotion or retirement is the engine's lane, D-1, and a chamber spec ruling an engine field's fate
+would exceed standing. STRENGTHENED after ruling by REVIEWED-107 §2: the vocabulary is undefined — three tokens in use,
+no definition in either repo — which is a cleaner ground than the one given here. Recorded as owed and unowned: SHA-STALE
+is now a fourth undefined token and no open item covers defining the vocabulary.
+
+5. Q3 — REVISED against my own lean, and the revision is the more useful half. The enumeration is not incomplete but NOT
+COMPLETABLE: the runbook's scope_note sets membership as any repo the engine manifest binds, and the runbook's own list
+was found short by its own grep on 2026-07-19. So the spec is authoritative for SEMANTICS and the runbook's grep for
+COMPLETENESS — two claims, two homes, not the fault condition 1 forbids. My original 'single enumerative authority' is
+withdrawn: it would have demoted the only instrument that has ever caught a missing surface.
+
+6. Q4 — refinement of the hash-locality principle's third instance, not a fourth. Same referent class, same home, finer
+granularity, and REVIEWED-53's individuating reason (files that move together) covers it. The ratified sentence stands
+untouched. TEST RECORDED, because the count is maintained by ruling and went 2 to 3 through REVIEWED-53: a future surface
+introducing a NEW HOME amends it; one refining an existing home does not.
+
+7. Q5 / Q6 — RENAME, not rescope in place. The voice_manifest incident that created this entry was a name inviting a
+wrong generalization; rescoping with a scope: field would be the same remedy a second time in the same block for the same
+failure mode. The ratified principle is NOT amended: L19 and L39-40 update as mechanical referring-name edits under
+REVIEWED-53's own lane rule for machine-data, with both reading grains preserved at the new name. REVIEWED.md L471 is not
+edited — a ruling records what it ruled. Name: canonical_binding, NOT canonical_binding_surface, which contains
+binding_surface and would have made the runbook's own key un-greppable through the instrument built to prevent that.
+Completion control required both directions: before, the search finds the known occurrences; after, zero hits on the old
+name outside REVIEWED.md, excluded BY NAME in the command.
+
+8. SUBSTRATE VERIFICATION. Both supplied files matched their declared shas and line counts, so the anchors held.
+Condition 2 leg (a) verified without the runbook — a key named engine already housed chamber artifacts in the executor's
+own draft. Leg (b) verified: catalogue.yaml at runbook L251. Two findings the package did not carry: the manifest binds
+THREE repos, not two (five after-the-reply sources are ARC, hard-coded to chamber paths in the draft), and R0 §4 L223-225
+is binary against §3 L180's collapse prohibition — split out as PENDING-127, since the chamber requirement was unmeetable
+while it stood. Now closed as REVIEWED-109, which found the code unary and 261 of 261 regions promoted under a hardcoded
+date.
+
+9. TWO SURFACES ADDED AFTER VERIFICATION. chamber-catalogue, and engine-sidecar-region — R0 §3 rules the two per-region
+gaps are one mechanism with two call sites, so enumerating only the reading index would have hard-coded that divergence
+into declared data.
+
+10. SPLITS AND COUPLING. PENDING-125 (Mauss, condition 5a) — a live false claim 53 days old must not die inside a
+deferred PROPOSAL; now REVIEWED-107(a), with (b) open. PENDING-127 — now REVIEWED-109. PENDING-128 — REVIEWED-53's
+deferred option (c), recorded as live again on the reasoning that carried Q6 and to be RULED jointly with this item. Its
+landing was made conditional rather than fixed: PENDING-127 clearing selects a single commit for both, and REVIEWED-109
+has cleared it, so the single commit obtains.
+
+11. SUPERSEDED EXAMPLE, flagged so it is not inherited. The item's design question cites Alexander's five stale
+front_matter anchors; REVIEWED-109 §4 records they were partitioned out on 2026-08-07 and that Alexander is 261/261 with
+zero unverified. The hazard is confirmed and larger, not weakened. Cite the 261/261 unconditional promotion.
+
+12. ON THE GROUNDING PASS. Five omissions across three passes, every substantive one understating the gap the executor
+was arguing for — the executor's own reading, accepted: passages stating the problem were quoted, passages stating its
+extent were skipped, and four of five were extent-passages. Not selective; systematic in kind.
+
+13. THE VERBATIM FILED RULING IS SUPERSEDED IN FOUR PLACES AND CARRIES NO MARKER. engine-source-binding-surfaces-JURIST-RULING-2026-08-08.md is the first pass only: its Stores line says the three files were NOT read and Parts I.1-I.4 are testimony (they were later read and sha-matched); its Q3 under 2(i) asserts the 'single enumerative authority' holding this entry withdraws at point 5; its condition 2 leaves the branch open (the steward chose (i)); and its §4 cites Alexander's five front_matter anchors, partitioned out 2026-08-07 per REVIEWED-109 §4. It is also the only verbatim jurist text in chamber docs/, and it contains no Q5 or Q6 — the rename holding PENDING-128 stands on exists there not at all. CONDITION: a supersession header at its top pointing to this entry, and to the two later passes, before either is relied on again. The document
+is not edited below the header — a filed ruling records what was ruled when.
+
+**If AUTHORIZED:** Conditions 1-4 stand as discharged in the Amendment 3 §D redraft, which is filed and awaiting the
+placement gate — that gate is NOT granted by this entry and requires the redrafted canonical_binding block itself, not
+its description. Rule jointly with PENDING-128; land in one commit per point 10. Tag with REVIEWED-110.
+
+## REVIEWED-111 — PENDING-128 — REVIEWED-53's deferred option (c): kill the manifest shared word
+**Date:** 2026-08-08
+**Decision:** DESIGN GATE PASSED on option (a); (c) rejected. Ruled jointly with PENDING-121 per §Coupling. The
+same-commit requirement is NARROWED, not adopted as filed. Three conditions.
+**Stores:** PENDING-128, PENDING-121 (incl. Amendments 1-3), REVIEWED-53, REVIEWED-107, REVIEWED-109 verbatim via the
+governance tools; graduation-spec.yaml L1-60 via governance_read. The shared-name collision log is reachable by no key
+and is executor testimony; it carries no weight here. Numbering assumes this follows REVIEWED-110 (PENDING-121).
+
+1. 1. RULED TOGETHER: YES, and the ground is REVIEWED-53's own text, not an unplaced holding. REVIEWED-53 judged (c) doctrinally complete and deferred it on occasion — 'out of scope for a doc-gap patch' — and PENDING-121 is a PROPOSAL that opens the same block deliberately, so the occasion has arrived. The parallel rename ruling is REVIEWED-110 §7; cite that, not 'PENDING-121's
+Q6', which appears in no placed or verbatim-filed record. Renaming one key because names must not mislead, while leaving its neighbour in the same block warned-around on the same ground, is incoherent. Neither is ruled without the other. ORDERING CONSEQUENCE: REVIEWED-110 must be placed before or with this entry, or point 1 cites forward into nothing.
+
+2. SAME COMMIT: CONDITIONALLY, AND NARROWED. The conditional filed against PENDING-127 has resolved — REVIEWED-109
+cleared it — so a single commit obtains. But PENDING-121 lands in TWO places: the mechanism in graduation-spec.yaml
+declared data, and the requirement in the constitution, MINOR bump by supersession. PENDING-128 is pure machine-data.
+Read as binding 128 to all of 121's landing, a machine-data rename would ride inside a constitutional supersession, and
+reverting the requirement would revert the rename — the revertability cost the conditional existed to avoid, returning
+through the door the blockage just left. RULED: the coupling binds PENDING-128 to PENDING-121's DECLARED-DATA landing —
+the layers: block commit — and not to its constitutional landing. §Coupling's own reason (same block, L19 cross-references
+L20 by name) supports exactly this scope and no more.
+
+3. OPTION (a) PASSES. Rename kills the collision. The executor's ground for recommending (a) and not (c) is affirmed:
+retiring a ratified safeguard should be its own decision with its own evidence, not a tidy-up riding on a rename.
+
+4. OPTION (c) REJECTED, and REVIEWED-107 strengthens the rejection. Retiring the warning rests on 'the name is now
+unambiguous', which is the assumption whose failure created this entry. REVIEWED-53 chose two reading grains
+deliberately. Both stay. Retiring a reading grain in a corpus just shown to mint undefined tokens is the wrong direction.
+
+5. CONDITION 1 — THE WARNING'S TEXT IS REWRITTEN, NOT MERELY KEPT. The item says keep the warning and does not notice
+that the rename changes what the warning is about. L19 currently warns that the word manifest names two different engine
+objects; after the rename the chamber side no longer carries that word, so the warning as written describes a collision
+that no longer exists at the site where it is printed. A kept-verbatim warning would be a stale safeguard — the shape
+this register keeps ruling against, arriving through a change made to improve clarity. Rewrite both grains to say what
+they now need to say: that the engine's SOURCE manifest binds by hash and is a different object, and that a reader
+arriving from an older citation of voice_manifest has reached the right entry. Preserving a safeguard means preserving
+its function, not its bytes.
+
+6. CONDITION 2 — THE NEW TERM IS DEFINED WHERE IT IS INTRODUCED, IN THE SAME ACT. REVIEWED-107 §2 found
+reading_index_status's vocabulary undefined — three tokens in use, no definition in either repo — and a fourth minted to
+state a truth none of the three could. That is a demonstrated corpus tendency to introduce terms without definitions and
+notice later. voice_personification is drawn from the entry's own prose; if personification is undefined at its site, the
+rename trades a documented collision for an undefined term, which is worse than the status quo, since the collision at
+least carried a warning. The definition goes in the rewritten grains of condition 1 — one act, not two.
+
+7. CONDITION 3 — THE COMPLETION CONTROL RUNS IN ONE INVOCATION, WITH A POSITIVE CONTROL. The item requires zero hits on
+the old name outside REVIEWED.md, and separately that the L19-L20 cross-reference still resolves in both directions. Run
+apart, the first is satisfiable by DELETING the cross-reference: the negative result passes precisely because the subject
+was removed. RULED: both run together, with resolves-at-new-names serving as the positive control for the zero-hits
+check. This hole opens only under a single commit, where the cross-reference becomes rewritable on both sides at once.
+
+8. NAME. voice_personification passes the substring test against binding_surface-class hazards. Recorded as neutral, not
+an improvement: graduation-spec.yaml carries voice as a frontmatter optional field in the same file, so a search for
+voice cannot isolate the frontmatter field from the layer key, before or after. The item's claim to be 'the first
+retired rather than warned around' should not be read as clearing the file of voice-family entanglement.
+
+9. UNVERIFIED, and carrying no weight. The shared-name collision log, and therefore 'the ninth such case'. Reachable by
+no governance_read key. The ground for (a) is REVIEWED-53's own text, which was read.
+
+10. NOT RULED HERE. The census of the other eight collisions, which the item correctly declines to propose. Whether
+REVIEWED.md L471 is touched — it is not; a ruling records what it ruled.
+
+**If AUTHORIZED:** Land option (a) with conditions 1-3, in PENDING-121's declared-data commit per point 2, not in its
+constitutional supersession. Tag with REVIEWED-111.

scripts/governance-drift-check.py

@@ -335,6 +335,21 @@ def trigger_fired(d: dict) -> bool | None:
 
 
 _scan_targets = [(r, "*/docs/**/*.md") for r in SCAN_ROOTS] + EXTRA_SCAN_GLOBS
+
+# PENDING-118, and the item's own option (1) is REFUTED by building it.
+# The item said "widen the scan to ~/PENDING-archive.md — the checker already parses
+# that exact format." It does not. The structured marker is an HTML comment
+# <!-- DEFERRED-DECISION: slug ... -->, and there are ZERO of those in PENDING.md or in
+# the archive. Their deferrals are PROSE — measured 2026-08-08: 53 and 26 occurrences of
+# "defer*". Widening alone would scan two more files, find nothing, and report clean:
+# a silent net built to close a blind spot, which is the failure this item describes.
+#
+# So the widening ships WITH its own honest limit. Structured blocks anywhere are now
+# found; prose deferrals are COUNTED and reported as un-machine-readable, never as
+# absent. Counting is not classifying — how many of those conditions have fired is a
+# READING task, and it is reported as owed rather than silently skipped.
+_scan_targets += [(HOME / "dotfiles", "PENDING.md"), (HOME / "dotfiles", "PENDING-archive.md")]
+PROSE_DEFERRAL_RE = re.compile(r"defer(?:red|ral)", re.I)
 _seen_files: set = set()
 for root, pattern in _scan_targets:
     if not root.is_dir():
@@ -381,6 +396,17 @@ control("trigger evaluator does NOT fire on an unmet condition  "
         _p_wait and trigger_fired(_p_wait[0]) is False)
 control("deferred-decision scan surface is reachable",
         any(r.is_dir() for r in SCAN_ROOTS))
+# Prose-deferral census: counted, never classified, and never read as absence.
+prose_counts = {}
+for _f in (HOME / "dotfiles" / "PENDING.md", HOME / "dotfiles" / "PENDING-archive.md"):
+    if _f.exists():
+        prose_counts[_f.name] = len(PROSE_DEFERRAL_RE.findall(_f.read_text(errors="replace")))
+control("prose-deferral counter finds a known-present phrase",
+        PROSE_DEFERRAL_RE.search("this was deferred pending recurrence") is not None)
+control("prose-deferral counter does not fire on unrelated text",
+        PROSE_DEFERRAL_RE.search("the fleet is green") is None)
+control("register files are inside the widened deferral scan",
+        any(str(t[1]).endswith("PENDING-archive.md") for t in _scan_targets))
 
 
 # ------------------------------------------ 9. built without a ruling in the register
@@ -532,5 +558,11 @@ if deferrals:
         waiting = len(deferrals) - len(manual)
         print(f"✓ deferred decisions: {len(deferrals)} tracked, none due "
               f"({waiting} checkable, {len(manual)} manual-only)")
+    if prose_counts:
+        _tot = sum(prose_counts.values())
+        print(f"  ⚠ plus {_tot} PROSE deferral mention(s) in the register "
+              f"({', '.join(f'{k} {v}' for k, v in prose_counts.items())}) — these carry no")
+        print("    DEFERRED-DECISION block, so NO trigger is machine-checkable for any of them.")
+        print("    Counted, not classified. Whether any condition has fired is unestablished.")
 
 sys.exit(0)