@event4u/agent-config 14.12.0 → 14.13.0

package/src/config/metric-registry.yml

@@ -0,0 +1,124 @@
+# Metric registry — consumer before producer
+#
+# `road-to-experience-loop-broadening` step 0.3. Every metric this package
+# produces declares three things before it lands:
+#
+#   consumer  — WHO reads it. A person, a gate, a report, a decision record.
+#   decision  — WHAT it changes. The decision it feeds, in one falsifiable line.
+#   absent    — WHAT FAILS without it. If nothing does, the metric is decoration.
+#
+# The rule the roadmap states, verbatim: *"A metric with no consumer is telemetry
+# decoration and should not land."* The tree's own worked example is the 0.27 %
+# dispatch capture figure — a collector whose consumer arrived after the data
+# did, so the number existed for months before anyone could say what it changed.
+#
+# ## What is a metric, for this file
+#
+# A number this package PRODUCES and keeps: a count in a report, a rate in a
+# ledger, a gauge in a gate's output. Not a threshold (that is a decision), not
+# a test count (that is a run), and not a number quoted from somewhere else
+# (that is a citation, and `docs/CLAIMS.md` governs those).
+#
+# ## The three fields are prose, and the lint checks shape rather than truth
+#
+# `lint_metric_consumers` refuses an entry with a missing, empty, or
+# boilerplate field. It cannot tell a real consumer from a plausible sentence —
+# that is a review judgement — so the honest claim is that the lint makes the
+# OMISSION impossible, not that it makes the answer true. Stated here rather
+# than discovered by someone who trusted it further than it goes.
+
+metrics:
+  - id: attribution-shape-block-count
+    producer: src/scripts/check_no_external_sources.ts
+    consumer: >-
+      the `check_no_external_sources:shape-block` ratchet, and the maintainer
+      reading its 56-day expiry.
+    decision: >-
+      whether a change that adds attribution-shaped text may merge. The ratchet
+      fails the build when the count rises above its baseline.
+    absent: >-
+      new attribution-shaped text lands unnoticed and the debt grows silently;
+      the deny-list half would still pass, because it only catches names
+      somebody already wrote down.
+
+  - id: skip-paths-estate
+    producer: src/scripts/check_suppression_hygiene.ts
+    consumer: the shrink-only suppression ratchet, and a reviewer reading its per-entry ledger.
+    decision: >-
+      whether a new scan exemption may be added. The count may only fall, so an
+      addition must be paired with a removal or an argued exception.
+    absent: >-
+      exemptions accumulate one defensible entry at a time until the gate covers
+      nothing, which is the failure mode the per-entry ledger exists to make
+      visible.
+
+  - id: gate-scanned-count
+    producer: src/scripts/_lib/scan_scope.ts (`reportScanned`, emitted by each gate)
+    consumer: src/scripts/check_gate_coverage.ts, against the per-gate `min_scanned` floor.
+    decision: >-
+      whether a gate's scan scope is still alive. A count below its floor fails
+      the build.
+    absent: >-
+      a gate whose root moved reports success over an empty corpus — "a gate
+      that scans nothing exits green", which is the specific way these gates
+      fail rather than a general worry.
+
+  - id: unhardened-gate-population
+    producer: src/scripts/check_gate_coverage.ts (`list_unhardened_gates`)
+    consumer: the `gate-hardening:unhardened-scan-scope` ratchet.
+    decision: whether a newly added gate may merge without publishing a scan count.
+    absent: >-
+      new gates arrive without scope assertions and the hardening effort silently
+      stops spreading; the existing gates keep their floors, so nothing else
+      would report the regression.
+
+  - id: registered-non-adopters
+    producer: src/scripts/check_gate_coverage.ts
+    consumer: the `gate-self-test:registered-non-adopters` shrink-only ratchet.
+    decision: >-
+      whether a newly registered gate may merge without a `--self-test` or an
+      explicit exemption marker.
+    absent: >-
+      a registered gate can prove it READ something (its `scanned:` floor) and
+      never that the reading changes its verdict, which is the difference
+      between a gate and a log line.
+
+  - id: always-rule-budget-chars
+    producer: src/scripts/check_always_budget.ts
+    consumer: the always-loaded budget ratchet, and any author adding a kernel rule.
+    decision: >-
+      whether a rule may be added to, or grown inside, the always-loaded
+      prefix. The cap is a ratchet that may only move down.
+    absent: >-
+      the standing context grows one justified rule at a time and every session
+      on every host pays for it, with no single change large enough to argue
+      about.
+
+  - id: council-quorum-present
+    producer: src/scripts/council_cli.ts
+    consumer: the human or agent reading a council verdict, and the release gate that holds on INCONCLUSIVE.
+    decision: >-
+      whether a council answer counts as convergence or as a degraded single
+      seat. A verdict with one of two seats present is recorded as DEGRADED and
+      is not convergence.
+    absent: >-
+      a one-seat answer reads as a council decision, which is exactly the
+      over-claim the quorum line exists to prevent.
+
+  - id: repeated-failure-rate
+    producer: src/scripts/_lib/repeated_failure.ts
+    consumer: >-
+      the human or CI reading the per-asset experience report, and the
+      pre-registered paired question in
+      `agents/evidence/experience-loop-prereg.md` that decides whether the
+      broadened loop is kept, narrowed, or dropped.
+    decision: >-
+      whether widening the learning loop changed anything measurable. A
+      reproducible fall in the rate at held quality and non-increased cost keeps
+      the loop; a rise, or no movement, is recorded as the pre-registered
+      negative and the phases resting on it are not built further.
+    absent: >-
+      the loop's own effect is unmeasured, so every later phase rests on the
+      assumption that widening helped -- which is the assumption the whole
+      roadmap exists to test rather than to make. Without it the estate grows
+      and nothing can say whether it should have.

package/src/config/pack-size-budget.json

@@ -1,21 +1,21 @@
 {
-    "_comment": "Payload budget for the published npm tarball (road-to-zero-ceremony-install Phase 4). SURFACE LABEL, added 2026-08-24 (2.3): every packed_size_mb figure in this file and in every baseline_note_* below was measured on the UNBUILT tree under measurement_conditions. The BUILT surface a consumer installs is larger and is recorded separately in built_surface_measurement_2026_08_24 \u2014 do not compare a note's figure against a built one. MEASURED, NEVER INVENTED \u2014 the numbers below come from `npm pack --dry-run --json --ignore-scripts` on the tracked tree at 9.9.0 (2026-07-31). Two metrics, both genuinely ungated before this file: COMPRESSED tarball size, and PER-SKILL share of the skills subtree. The UNPACKED size is deliberately absent, and the reason CHANGED on 2026-08-04 without this sentence being updated \u2014 corrected 2026-08-24. It used to read that unpacked size is 'already owned by src/config/evaluator-budgets.json (unpacked_size_mb)'. That key was REMOVED by maintainer decision at the 9.17.0 release (`evaluator-budgets.json` -> `removed_2026_08_04`): the gate turned every legitimate payload addition into a blocked release (28.22 measured against max 28, all functional gates green) and the maintainer judged tarball size not worth gating. So the built artifact's unpacked size is MEASURED as evidence by `evaluator_umbrella.sh` and is deliberately NOT budget-gated. Reinstating `unpacked_size_mb` reverses a recorded maintainer release decision and requires owner review under `decision-revisit-gate` \u2014 it is not a wiring fix. The original point stands on its own terms: a second number for one lever is how three conflicting figures start circulating.",
+    "_comment": "Payload budget for the published npm tarball (road-to-zero-ceremony-install Phase 4). SURFACE LABEL, added 2026-08-24 (2.3): every packed_size_mb figure in this file and in every baseline_note_* below was measured on the UNBUILT tree under measurement_conditions. The BUILT surface a consumer installs is larger and is recorded separately in built_surface_measurement_2026_08_24 — do not compare a note's figure against a built one. MEASURED, NEVER INVENTED — the numbers below come from `npm pack --dry-run --json --ignore-scripts` on the tracked tree at 9.9.0 (2026-07-31). Two metrics, both genuinely ungated before this file: COMPRESSED tarball size, and PER-SKILL share of the skills subtree. The UNPACKED size is deliberately absent, and the reason CHANGED on 2026-08-04 without this sentence being updated — corrected 2026-08-24. It used to read that unpacked size is 'already owned by src/config/evaluator-budgets.json (unpacked_size_mb)'. That key was REMOVED by maintainer decision at the 9.17.0 release (`evaluator-budgets.json` -> `removed_2026_08_04`): the gate turned every legitimate payload addition into a blocked release (28.22 measured against max 28, all functional gates green) and the maintainer judged tarball size not worth gating. So the built artifact's unpacked size is MEASURED as evidence by `evaluator_umbrella.sh` and is deliberately NOT budget-gated. Reinstating `unpacked_size_mb` reverses a recorded maintainer release decision and requires owner review under `decision-revisit-gate` — it is not a wiring fix. The original point stands on its own terms: a second number for one lever is how three conflicting figures start circulating.",
     "schema_version": 1,
     "owner": "maintainer",
     "review_by": "2027-07-31",
     "regression_pct": 10,
-    "measurement_conditions": "npm pack --dry-run --json --ignore-scripts, run at the repo root on a clean checkout. `--ignore-scripts` means NO build runs, so dist/cli, dist/ui, dist/mcp and dist/hooks are absent and the figure covers only the TRACKED payload (src/scripts/, dist/agent-src/, dist/install/, docs/guidelines/, templates). That is deliberate: this gate is always-on for every PR, and the tracked content tree is exactly the surface that grows by accretion. The built artifact is ~5.4 MB larger and is NOT gated \u2014 corrected 2026-08-24. This sentence named `evaluator-budgets.unpacked_size_mb` as a release-time gate; that key was removed by maintainer decision on 2026-08-04 (see `_comment`). `evaluator_umbrella.sh` still measures and records the figure as evidence, so the number exists and nothing enforces it. That is the honest state and it is deliberate.",
+    "measurement_conditions": "npm pack --dry-run --json --ignore-scripts, run at the repo root on a clean checkout. `--ignore-scripts` means NO build runs, so dist/cli, dist/ui, dist/mcp and dist/hooks are absent and the figure covers only the TRACKED payload (src/scripts/, dist/agent-src/, dist/install/, docs/guidelines/, templates). That is deliberate: this gate is always-on for every PR, and the tracked content tree is exactly the surface that grows by accretion. The built artifact is ~5.4 MB larger and is NOT gated — corrected 2026-08-24. This sentence named `evaluator-budgets.unpacked_size_mb` as a release-time gate; that key was removed by maintainer decision on 2026-08-04 (see `_comment`). `evaluator_umbrella.sh` still measures and records the figure as evidence, so the number exists and nothing enforces it. That is the honest state and it is deliberate.",
     "budgets": {
         "packed_size_mb": {
             "max": 9.1,
             "last_measured": 8.475,
-            "baseline_note_2026_08_23": "8.4 \u2192 9.2 re-baselined 2026-08-23. RATCHET RESET, not the config bent around a failing check: the cap was exhausted at TRUNK level before the tripping branch existed. MEASURED on clean clones under the documented conditions \u2014 origin/main at f16c7d9df reads 8.3911, i.e. 8.9 KB under its own cap; PR #1597 reads 8.4083 local / 8.416 CI, +17.2 KB across 4 new pack entries of functional source its roadmap deliberately adds. With 8.9 KB of headroom ANY commit reds this gate, which is the failure the 2026-08-20 note predicted in writing, arriving a second time three days later. Max = 8.4083 \u00d7 1.095, derived from this file's own regression_pct: 10 rather than from the tripping measurement; last_measured records CI's 8.416, the enforcing environment and the higher of the disagreeing pair. GIVEN UP: ~0.79 MB of unpoliced accretion \u2014 this gate catches a step change in payload, which regression_pct still catches at any cap, and it cannot police 17 KB increments. REJECTED, measured rather than argued: payload reduction inside the PR \u2014 the whitelist is already tight (docs/guidelines/ plus three named contracts), tests/ is 6 KB and agents/ 39 KB uncompressed, so excluding both saves ~13 KB compressed and does not clear even this one PR. The structural surface the 2026-08-20 note named (src/scripts/ at 14.9 MB unpacked, src/agent-src/ duplicating dist/agent-src/) is still unexamined and still the open debt this raise defers. Revisit-if: a single PR moves the figure by more than regression_pct, or the drain ends and the cap can ratchet DOWN from a re-measured trunk \u2014 never up again from a fresh trip. DECIDED BY: maintainer, numbered-options option 1, after the alternative above was measured and the exhaustion was reported rather than the cap raised silently.",
-            "method": "npm pack --dry-run --json --ignore-scripts \u2192 dist.size / 1e6 (the compressed tarball a consumer actually downloads)",
-            "baseline_note_2026_08_24": "8.4 \u2192 9.1 re-baselined 2026-08-24, and this is a RATCHET RESET rather than the config bent around a failing check. The discriminator this file demands, answered with a measurement rather than an argument: THE TRUNK WAS ALREADY AT THE CEILING BEFORE THE CHANGE THAT TRIPPED IT EXISTED. `origin/main` at f16c7d9df measures 8.3911 under the documented conditions \u2014 8.9 KB of headroom against an 8.4 cap, i.e. 0.1 %. An addition of any size crosses it. That is verbatim the failure mode the 2026-08-20 note predicted in writing \u2014 'a cap pinned just above the measurement reds on the next commit and teaches the reader to raise it again' \u2014 arriving on schedule for the SECOND time, which is why the correction here is the headroom and not the payload. TRIP CAUSE: merging `origin/main` into `drain/chained-clip-continuity` (PR #1598). The branch adds ~2,600 net lines inside packed paths \u2014 `src/scripts/lint_adapter_tier.ts` (662 lines; it is the gate that verifies the generated tier table in `docs/contracts/provider-lifecycle.md`, so it must ship), the ai-video adapter / seam-plan / stitch shell libraries, and the `model-capabilities/*.json` files read at runtime \u2014 landing 39 KB over the cap. NEITHER SIDE WAS OVER ALONE; together they cross, which is the shape a merge produces and no per-branch check can see. MEASURED, on two clean detached worktrees with unbuilt `dist/` per `measurement_conditions` (so dist/cli, dist/ui, dist/mcp and dist/hooks absent): main 8.3911 (2,746 entries) \u00b7 branch 8.4305 (2,753 entries) \u00b7 CI 8.439 (Consistency job 97266868959). `last_measured` records CI's 8.439 \u2014 the enforcing environment and the higher of the disagreeing pair, per the rule the 2026-08-10 and 2026-08-20 notes both applied to their own; the ~8 KB CI-over-local gap matches the pattern both recorded. The 10 % regression rule did NOT trip (threshold 8.586 against the prior `last_measured` 7.805); only the absolute cap did. Max set to 9.1, ~7.8 % headroom over 8.439, which is the ~8 % the 5.903/6.4, 7.238/7.8 and 7.805/8.4 pairs all carried: deliberately not a number that merely clears. ALTERNATIVE EXAMINED AND NOT TAKEN, measured rather than argued, and its verdict left HONESTLY OPEN: excluding `src/scripts/ai_council/` \u2014 the tree the 2026-08-20 note names as the one whose consumer need is 'unverified', 0.903 MB as measured here \u2014 saves 270 KB compressed (8.4305 \u2192 8.1607, under the OLD cap with 239 KB to spare). It is not taken here because whether it is SAFE is still undecided, and this note deliberately does not decide it. `src/server/routes/wizard.ts:48` imports `apiOnQuotaView` from `scripts/ai_council/transport_resolver.js`, and `council:status` is a registered verb in `src/cli/registry.ts` \u2014 but the specifier ends in `.js` and therefore cannot settle whether it resolves to the TS source under `src/scripts/` or to built output under `dist/` in a published install. An earlier draft of this note asserted that the exclusion 'breaks the wizard quota view in every consumer install'; that was an over-claim read off the specifier, and it is retracted here rather than quietly dropped. What IS established: `src/cli/python/workspace_hosts.ts:180-191` records a real `ERR_MODULE_NOT_FOUND` in a global install caused by a shipped module importing from a directory absent from `files[]`, and `src/server/routes/wizard.ts:136-137` invokes `bash src/scripts/install_anthropic_key.sh` BY PATH at runtime. So a blanket trim of `src/scripts/` is unsafe, while the ai_council-specific verdict needs a global-install smoke test rather than a reading of an import. THE DEBT IS NOW WRITTEN DOWN, which is the one thing the 2026-08-20 note promised and did not deliver: it said 'this raise buys the time to write one [a roadmap]', and no roadmap was ever written \u2014 a grep of `agents/roadmaps/` (active, `stubs/`, `later/`, `archive/`) finds no payload-reduction plan. `agents/roadmaps/archive/road-to-npm-payload-reduction.md` lands in this same change and carries the measured split: `src/scripts/` is 16.384 MB of the 28.604 MB unpacked payload (57.3 %, 1,158 files) and is the actual growth surface; `dist/agent-src/` is 8.040 MB; `src/agent-src/` is 1.347 MB and partly duplicates it, unmeasured until that roadmap's Phase 1.3. Its Phase 2.3 lowers this cap again, so `max: 9.1` still standing later is the visible signal the plan did not finish. [AMENDED 2026-08-24 by the run that finished it, because that sentence would otherwise mislead in the opposite direction: the plan DID finish, and 9.1 is now its RESULT rather than its deferral signal \u2014 see baseline_note_2026_08_24_lowered. The roadmap path in this sentence was rewritten from agents/roadmaps/ to agents/roadmaps/archive/ by the archival ref-migration, so a dated note now points at a path that did not exist on its own date; the migration is kept because the reference must resolve, and it is flagged here rather than left to read as if the file had always been archived.] A raise that defers the same debt a fourth time with no plan is a formality, so this one carries the plan. DECIDED BY: maintainer, on the PR that tripped it, chosen from numbered options carrying the measurements above rather than applied to make a red go away \u2014 and the write itself was performed by the maintainer, because the block-config-weakening guard refused four agent attempts at this file. That is the standing this file's `owner: maintainer` requires, enforced rather than asserted. MET A SECOND, INDEPENDENT RESET IN A MERGE, 2026-08-24: the drain's council-evidence-integrity branch reached the same diagnosis from the same measurement (main at 8.3911, 8.9 KB of headroom) and set 9.2 instead of 9.1, landing on trunk first. The HIGHER cap prevails and BOTH derivations are kept, because two independent readings converging on one trip cause is the strongest evidence this entry has ever carried \u2014 and dropping one of them would leave the record looking like a single unreviewed judgement. `last_measured` is re-measured on the merged tree rather than taken from either side, per the method every prior entry here prescribes.",
-            "baseline_note_2026_08_20": "7.8 \u2192 8.4 re-baselined 2026-08-20, and this is a RATCHET RESET rather than the config bent around a failing check \u2014 the discriminator the 2026-08-10 note demands, answered here. TRIP CAUSE: nothing was added. `origin/main` at cfd8426d9 measures 7.7916 under the documented conditions, and CI reads ~7 KB above the local figure, so trunk itself already stood ~8 KB below a 7.8 cap before the release branch existed; the 14.6.0 CHANGELOG entry (6.3 KB \u2014 a file on the release-PR shape allowlist, i.e. the release's own permitted payload) is what crossed it, at 7.7979 local / 7.805 CI. That is precisely the failure mode the 2026-08-10 note predicted in writing \u2014 'a cap pinned just above the measurement reds on the next commit and teaches the reader to raise it again' \u2014 arriving on schedule, which is why the correction here is the headroom and not the payload. MEASURED, on a clean clone of each tree with the method `measurement_conditions` documents (unbuilt dist/, so dist/cli, dist/ui, dist/mcp and dist/hooks absent): main 7.7916 \u00b7 release/14.6.0 7.7979 \u00b7 CI 7.805 (Consistency job 96239721013). `last_measured` records CI's 7.805 \u2014 the enforcing environment and the higher of the disagreeing pair, per the same rule the 2026-08-10 note applied to its own. Max set to 8.4, ~7.6 % headroom over 7.805, which is the ~8 % both the 5.903/6.4 and 7.238/7.8 pairs carried: deliberately not a number that merely clears. ALTERNATIVE EXAMINED AND REJECTED, measured rather than argued: forcing the CHANGELOG era split early \u2014 the one payload reduction that IS on the release-PR allowlist, so it could have landed without touching main at all \u2014 moves the 144-line 14.5.0 section (7,266 bytes) into `docs/archive/`, which sits outside `files[]`. It saves 1.0 KB compressed (7.7979 \u2192 7.7969). Prose compresses; a ~7 KB deficit cannot be closed by relocating 7 KB of prose. Recorded so it is not re-proposed as the cheap fix it looks like. NOT EXAMINED HERE, and named as the open question this raise defers rather than answers: `src/scripts/` is 14.9 MB of the 26.0 MB unpacked payload and is the actual growth surface, `src/agent-src/` partly duplicates `dist/agent-src/`, and whether a consumer needs the 846 KB `src/scripts/ai_council/` tree at all is unverified. Each is a structural change with its own blast radius, not a release-time edit \u2014 that belongs in a roadmap, and this raise buys the time to write one. A debt this note records, not one it discharges. DECIDED BY: maintainer, on the release that tripped it, in its own PR against main \u2014 `src/config/pack-size-budget.json` is not on the release-PR shape allowlist, so this cannot ride the release branch and is not smuggled into one.",
-            "baseline_note_2026_08_10": "6.9 \u2192 7.8 re-baselined 2026-08-10. WHICH OF THE TWO THIS IS, stated because the guard asks: a RATCHET RESET, not the config bent around a failing check \u2014 and the discriminator is that the payload is real and needed rather than that the check was inconvenient. The 44 KB is two source files a shipped instruction depends on, the cheaper alternative was examined and rejected on that ground (below), the new cap carries ~8% headroom over the higher of two disagreeing measurements rather than clearing by a hair, and the raise was put to the maintainer as a decision instead of applied to make a red go away. A reader who disagrees has everything needed to reverse it: the measurement, its method, the trip cause, and the rejected option. Maintainer-decided in the PR that tripped it, surfaced rather than smuggled \u2014 the same standing this file's `owner: maintainer` requires. MEASURED TWICE, and the two do not agree: CI reported 6.907 on this exact commit, a local run under the documented conditions reports 7.238. The 331 KB gap is UNEXPLAINED \u2014 the tree matches the pushed commit and only this file differs \u2014 so the cap is set over the HIGHER figure rather than the convenient one, and the discrepancy is recorded here as an open question instead of resolved by assumption. A third party reproducing this should expect either number and should not treat the lower one as authoritative. Separately, and this part IS understood: `npm pack --dry-run --ignore-scripts` in a worktree that has been BUILT reads 8.45, because the flag assumes dist/cli, dist/ui, dist/mcp and dist/hooks are absent and a built tree has them. That figure is not comparable to this budget and must never be written here \u2014 recorded because it is a trap the documented method invites. Trip cause: road-to-conformance-round5 Phase 3 added src/scripts/hooks/turn_end_gate_hook.ts (32 KB) and src/scripts/measure_turn_end_gate.ts (12 KB) in packed paths plus schema/template/contract growth, landing 7 KB over a 6.9 cap. Max set to 7.8 (~8% headroom over 7.238, and ~13% over CI's 6.907) for the reason the 2026-08-05 note already gives: a cap pinned just above the measurement reds on the next commit and teaches the reader to raise it again. ALTERNATIVE EXAMINED AND REJECTED, so it is not re-proposed: dropping measure_turn_end_gate.ts from the pack would have cleared the cap with room, but the settings description added in the same change instructs CONSUMERS to run that script against their own transcript store before arming the gate \u2014 excluding it would break a shipped instruction.",
-            "baseline_note": "5.903 \u2192 6.398 re-baselined 2026-08-05. This is a RE-BASELINE, not a carve-out for the change that tripped it: measured on the tracked tree, `origin/main` alone was already at 6.372 of a 6.4 cap, so 28 KB of headroom remained after months of accretion and the next legitimate addition of any size was going to cross it. The addition here is 26 KB \u2014 docs/contracts/settings-classes.md (the A/B/C fence, read at RUNTIME by the settings writer, so it must ship) plus src/server/io/yamlIO.ts (imported by that writer to preserve the user's comments). CI measures ~4 KB above the local figure because it regenerates dist/agent-src first. Max set to 6.9 rather than to a number that merely clears: a cap pinned 2 KB above the measurement reds on the next commit and teaches the reader to raise it again, which is how a budget becomes a formality. That restores the ~8% headroom the 5.903/6.4 pair originally carried. The `owner: maintainer` on this file stands \u2014 this raise is surfaced in the PR for that decision, not smuggled.",
-            "baseline_note_2026_08_24_lowered": "9.2 -> 9.1 LOWERED 2026-08-24 \u2014 the first DOWNWARD move in this file's history, and the reason it is possible is that road-to-npm-payload-reduction measured the lever the previous four raises kept deferring. Method, because the verdicts turn on it: a behavioural global-install harness (src/scripts/pack_install_smoke.ts) packs a real BUILT tarball, installs it into a throwaway prefix, and runs eleven consumer probes including the bash orchestrator FROM the installed tree, which is the only probe that exercises path-reachability rather than imports. Sensitivity was proven by sabotage before any verdict was trusted: excluding src/scripts/_lib/ produces ERR_MODULE_NOT_FOUND on four probes. WHAT THE BISECT FOUND, and it retires a suggestion this file has carried since 2026-08-20: EVERY named subtree SHIPS. src/scripts/ai_council/ \u2014 the 270 KB candidate the 2026-08-24 note left explicitly undecided \u2014 breaks council:status AND hooks:status with ERR_MODULE_NOT_FOUND. So does src/scripts/hooks/ (271.5 KB), src/scripts/_cli/ (204.1 KB), src/agent-src/ (101.8 KB), src/scripts/ai-video/ (95.4 KB, its shell entry point), src/scripts/mcp_server/ (53.1 KB). Two of those first read GREEN and flipped RED when the probe set was widened with mcp:setup and an ai-video shell check \u2014 recorded because it is the exact false-negative shape this file's own Risk 2 names, caught rather than shipped. WHAT IS ACTUALLY REMOVABLE is three FILE PATTERNS, not subtrees, and no prior note looked for them: dist/**/*.map (323.5 KB), dist/agent-src/skills/**/evals/** (56.2 KB), src/**/*.test.ts + *.spec.ts (22.6 KB) \u2014 402.9 KB on the built tarball, all eleven probes green together. MEASURED under this file's own documented conditions (npm pack --dry-run --ignore-scripts, clean unbuilt worktree, which is what check_pack_size reads): 8.5596 -> 8.4667 MB, 2,775 -> 2,609 entries, a 92.9 KB drop. The gap between 402.9 KB and 92.9 KB is not a discrepancy: most of the .map saving lives in dist/cli, dist/ui, dist/mcp and dist/hooks, which --ignore-scripts excludes by design. HEADROOM IS 7.4 %, NOT 8 %, and the number is stated exactly rather than rounded up to match the pattern: 9.1 over the CI-adjusted 8.475 (local 8.4667 plus the ~8 KB CI delta every prior note here records) is 7.37 %. The 5.903/6.4, 7.238/7.8, 7.805/8.4 and 8.439/9.1 pairs carried ~8 %; this one carries slightly less, which is a real if small increase in flake risk and is why the falsifier below names it. STILL OPEN, quantified for the first time: src/agent-src/ is 94.2 % byte-identical to dist/agent-src/ (284 of 297 files, 2,144,870 of 2,276,707 bytes, content-hash diff). Two prior notes called this 'partial'; it is near-total. It is NOT removable by exclusion \u2014 nine shipped scripts under src/scripts/ import through it (build_archive_index, check_estate_count, eval_ui_triviality, explain_run, lint_ui_stack_bundles, skill_usage_report, smoke_quickstart, telemetry_disclosure_hook, utilization_report) while dist/agent-src/ is the tree the installer deploys, so the two copies ship for different reasons. Repointing those imports at dist/ would make source depend on a generated projection, which src/rules/source-of-truth.md exists to prevent. AI council 2026-08-24, 2/2: that needs an ADR covering compiled script entry points and asset resolution, not another files[] experiment. Worth ~102 KB and deliberately not taken here. TRADE-OFF ACCEPTED, named rather than absorbed: excluding dist/**/*.map means published packages carry no source maps, so consumer stack traces show compiled names. GREEN here means 'no regression across the recorded eleven-probe matrix', never 'semantically equivalent'. REVISIT-IF: three clean CI runs measure above 9.1 (the method does not match the gate); or observed run-to-run variance exceeds the 7.4 % margin, making the ratchet noise rather than a floor; or the src/agent-src/ ADR lands and the cap can drop again. NEVER raise it again from a fresh trip without the discriminator the 2026-08-10 note demands. DECIDED BY: AI council 2026-08-24, 2/2 convergent (anthropic/claude-sonnet-4-5 + openai/codex-default, 2 rounds, blind peer review), on a maintainer delegation of owner-reserved blockers for an autonomous drain run. Both seats independently required the 7.4 % to be stated exactly and the GREEN to be scoped to the probe matrix."
+            "baseline_note_2026_08_23": "8.4 → 9.2 re-baselined 2026-08-23. RATCHET RESET, not the config bent around a failing check: the cap was exhausted at TRUNK level before the tripping branch existed. MEASURED on clean clones under the documented conditions — origin/main at f16c7d9df reads 8.3911, i.e. 8.9 KB under its own cap; PR #1597 reads 8.4083 local / 8.416 CI, +17.2 KB across 4 new pack entries of functional source its roadmap deliberately adds. With 8.9 KB of headroom ANY commit reds this gate, which is the failure the 2026-08-20 note predicted in writing, arriving a second time three days later. Max = 8.4083 × 1.095, derived from this file's own regression_pct: 10 rather than from the tripping measurement; last_measured records CI's 8.416, the enforcing environment and the higher of the disagreeing pair. GIVEN UP: ~0.79 MB of unpoliced accretion — this gate catches a step change in payload, which regression_pct still catches at any cap, and it cannot police 17 KB increments. REJECTED, measured rather than argued: payload reduction inside the PR — the whitelist is already tight (docs/guidelines/ plus three named contracts), tests/ is 6 KB and agents/ 39 KB uncompressed, so excluding both saves ~13 KB compressed and does not clear even this one PR. The structural surface the 2026-08-20 note named (src/scripts/ at 14.9 MB unpacked, src/agent-src/ duplicating dist/agent-src/) is still unexamined and still the open debt this raise defers. Revisit-if: a single PR moves the figure by more than regression_pct, or the drain ends and the cap can ratchet DOWN from a re-measured trunk — never up again from a fresh trip. DECIDED BY: maintainer, numbered-options option 1, after the alternative above was measured and the exhaustion was reported rather than the cap raised silently.",
+            "method": "npm pack --dry-run --json --ignore-scripts → dist.size / 1e6 (the compressed tarball a consumer actually downloads)",
+            "baseline_note_2026_08_24": "8.4 → 9.1 re-baselined 2026-08-24, and this is a RATCHET RESET rather than the config bent around a failing check. The discriminator this file demands, answered with a measurement rather than an argument: THE TRUNK WAS ALREADY AT THE CEILING BEFORE THE CHANGE THAT TRIPPED IT EXISTED. `origin/main` at f16c7d9df measures 8.3911 under the documented conditions — 8.9 KB of headroom against an 8.4 cap, i.e. 0.1 %. An addition of any size crosses it. That is verbatim the failure mode the 2026-08-20 note predicted in writing — 'a cap pinned just above the measurement reds on the next commit and teaches the reader to raise it again' — arriving on schedule for the SECOND time, which is why the correction here is the headroom and not the payload. TRIP CAUSE: merging `origin/main` into `drain/chained-clip-continuity` (PR #1598). The branch adds ~2,600 net lines inside packed paths — `src/scripts/lint_adapter_tier.ts` (662 lines; it is the gate that verifies the generated tier table in `docs/contracts/provider-lifecycle.md`, so it must ship), the ai-video adapter / seam-plan / stitch shell libraries, and the `model-capabilities/*.json` files read at runtime — landing 39 KB over the cap. NEITHER SIDE WAS OVER ALONE; together they cross, which is the shape a merge produces and no per-branch check can see. MEASURED, on two clean detached worktrees with unbuilt `dist/` per `measurement_conditions` (so dist/cli, dist/ui, dist/mcp and dist/hooks absent): main 8.3911 (2,746 entries) · branch 8.4305 (2,753 entries) · CI 8.439 (Consistency job 97266868959). `last_measured` records CI's 8.439 — the enforcing environment and the higher of the disagreeing pair, per the rule the 2026-08-10 and 2026-08-20 notes both applied to their own; the ~8 KB CI-over-local gap matches the pattern both recorded. The 10 % regression rule did NOT trip (threshold 8.586 against the prior `last_measured` 7.805); only the absolute cap did. Max set to 9.1, ~7.8 % headroom over 8.439, which is the ~8 % the 5.903/6.4, 7.238/7.8 and 7.805/8.4 pairs all carried: deliberately not a number that merely clears. ALTERNATIVE EXAMINED AND NOT TAKEN, measured rather than argued, and its verdict left HONESTLY OPEN: excluding `src/scripts/ai_council/` — the tree the 2026-08-20 note names as the one whose consumer need is 'unverified', 0.903 MB as measured here — saves 270 KB compressed (8.4305 → 8.1607, under the OLD cap with 239 KB to spare). It is not taken here because whether it is SAFE is still undecided, and this note deliberately does not decide it. `src/server/routes/wizard.ts:48` imports `apiOnQuotaView` from `scripts/ai_council/transport_resolver.js`, and `council:status` is a registered verb in `src/cli/registry.ts` — but the specifier ends in `.js` and therefore cannot settle whether it resolves to the TS source under `src/scripts/` or to built output under `dist/` in a published install. An earlier draft of this note asserted that the exclusion 'breaks the wizard quota view in every consumer install'; that was an over-claim read off the specifier, and it is retracted here rather than quietly dropped. What IS established: `src/cli/python/workspace_hosts.ts:180-191` records a real `ERR_MODULE_NOT_FOUND` in a global install caused by a shipped module importing from a directory absent from `files[]`, and `src/server/routes/wizard.ts:136-137` invokes `bash src/scripts/install_anthropic_key.sh` BY PATH at runtime. So a blanket trim of `src/scripts/` is unsafe, while the ai_council-specific verdict needs a global-install smoke test rather than a reading of an import. THE DEBT IS NOW WRITTEN DOWN, which is the one thing the 2026-08-20 note promised and did not deliver: it said 'this raise buys the time to write one [a roadmap]', and no roadmap was ever written — a grep of `agents/roadmaps/` (active, `stubs/`, `later/`, `archive/`) finds no payload-reduction plan. `agents/roadmaps/archive/road-to-npm-payload-reduction.md` lands in this same change and carries the measured split: `src/scripts/` is 16.384 MB of the 28.604 MB unpacked payload (57.3 %, 1,158 files) and is the actual growth surface; `dist/agent-src/` is 8.040 MB; `src/agent-src/` is 1.347 MB and partly duplicates it, unmeasured until that roadmap's Phase 1.3. Its Phase 2.3 lowers this cap again, so `max: 9.1` still standing later is the visible signal the plan did not finish. [AMENDED 2026-08-24 by the run that finished it, because that sentence would otherwise mislead in the opposite direction: the plan DID finish, and 9.1 is now its RESULT rather than its deferral signal — see baseline_note_2026_08_24_lowered. The roadmap path in this sentence was rewritten from agents/roadmaps/ to agents/roadmaps/archive/ by the archival ref-migration, so a dated note now points at a path that did not exist on its own date; the migration is kept because the reference must resolve, and it is flagged here rather than left to read as if the file had always been archived.] A raise that defers the same debt a fourth time with no plan is a formality, so this one carries the plan. DECIDED BY: maintainer, on the PR that tripped it, chosen from numbered options carrying the measurements above rather than applied to make a red go away — and the write itself was performed by the maintainer, because the block-config-weakening guard refused four agent attempts at this file. That is the standing this file's `owner: maintainer` requires, enforced rather than asserted. MET A SECOND, INDEPENDENT RESET IN A MERGE, 2026-08-24: the drain's council-evidence-integrity branch reached the same diagnosis from the same measurement (main at 8.3911, 8.9 KB of headroom) and set 9.2 instead of 9.1, landing on trunk first. The HIGHER cap prevails and BOTH derivations are kept, because two independent readings converging on one trip cause is the strongest evidence this entry has ever carried — and dropping one of them would leave the record looking like a single unreviewed judgement. `last_measured` is re-measured on the merged tree rather than taken from either side, per the method every prior entry here prescribes.",
+            "baseline_note_2026_08_20": "7.8 → 8.4 re-baselined 2026-08-20, and this is a RATCHET RESET rather than the config bent around a failing check — the discriminator the 2026-08-10 note demands, answered here. TRIP CAUSE: nothing was added. `origin/main` at cfd8426d9 measures 7.7916 under the documented conditions, and CI reads ~7 KB above the local figure, so trunk itself already stood ~8 KB below a 7.8 cap before the release branch existed; the 14.6.0 CHANGELOG entry (6.3 KB — a file on the release-PR shape allowlist, i.e. the release's own permitted payload) is what crossed it, at 7.7979 local / 7.805 CI. That is precisely the failure mode the 2026-08-10 note predicted in writing — 'a cap pinned just above the measurement reds on the next commit and teaches the reader to raise it again' — arriving on schedule, which is why the correction here is the headroom and not the payload. MEASURED, on a clean clone of each tree with the method `measurement_conditions` documents (unbuilt dist/, so dist/cli, dist/ui, dist/mcp and dist/hooks absent): main 7.7916 · release/14.6.0 7.7979 · CI 7.805 (Consistency job 96239721013). `last_measured` records CI's 7.805 — the enforcing environment and the higher of the disagreeing pair, per the same rule the 2026-08-10 note applied to its own. Max set to 8.4, ~7.6 % headroom over 7.805, which is the ~8 % both the 5.903/6.4 and 7.238/7.8 pairs carried: deliberately not a number that merely clears. ALTERNATIVE EXAMINED AND REJECTED, measured rather than argued: forcing the CHANGELOG era split early — the one payload reduction that IS on the release-PR allowlist, so it could have landed without touching main at all — moves the 144-line 14.5.0 section (7,266 bytes) into `docs/archive/`, which sits outside `files[]`. It saves 1.0 KB compressed (7.7979 → 7.7969). Prose compresses; a ~7 KB deficit cannot be closed by relocating 7 KB of prose. Recorded so it is not re-proposed as the cheap fix it looks like. NOT EXAMINED HERE, and named as the open question this raise defers rather than answers: `src/scripts/` is 14.9 MB of the 26.0 MB unpacked payload and is the actual growth surface, `src/agent-src/` partly duplicates `dist/agent-src/`, and whether a consumer needs the 846 KB `src/scripts/ai_council/` tree at all is unverified. Each is a structural change with its own blast radius, not a release-time edit — that belongs in a roadmap, and this raise buys the time to write one. A debt this note records, not one it discharges. DECIDED BY: maintainer, on the release that tripped it, in its own PR against main — `src/config/pack-size-budget.json` is not on the release-PR shape allowlist, so this cannot ride the release branch and is not smuggled into one.",
+            "baseline_note_2026_08_10": "6.9 → 7.8 re-baselined 2026-08-10. WHICH OF THE TWO THIS IS, stated because the guard asks: a RATCHET RESET, not the config bent around a failing check — and the discriminator is that the payload is real and needed rather than that the check was inconvenient. The 44 KB is two source files a shipped instruction depends on, the cheaper alternative was examined and rejected on that ground (below), the new cap carries ~8% headroom over the higher of two disagreeing measurements rather than clearing by a hair, and the raise was put to the maintainer as a decision instead of applied to make a red go away. A reader who disagrees has everything needed to reverse it: the measurement, its method, the trip cause, and the rejected option. Maintainer-decided in the PR that tripped it, surfaced rather than smuggled — the same standing this file's `owner: maintainer` requires. MEASURED TWICE, and the two do not agree: CI reported 6.907 on this exact commit, a local run under the documented conditions reports 7.238. The 331 KB gap is UNEXPLAINED — the tree matches the pushed commit and only this file differs — so the cap is set over the HIGHER figure rather than the convenient one, and the discrepancy is recorded here as an open question instead of resolved by assumption. A third party reproducing this should expect either number and should not treat the lower one as authoritative. Separately, and this part IS understood: `npm pack --dry-run --ignore-scripts` in a worktree that has been BUILT reads 8.45, because the flag assumes dist/cli, dist/ui, dist/mcp and dist/hooks are absent and a built tree has them. That figure is not comparable to this budget and must never be written here — recorded because it is a trap the documented method invites. Trip cause: road-to-conformance-round5 Phase 3 added src/scripts/hooks/turn_end_gate_hook.ts (32 KB) and src/scripts/measure_turn_end_gate.ts (12 KB) in packed paths plus schema/template/contract growth, landing 7 KB over a 6.9 cap. Max set to 7.8 (~8% headroom over 7.238, and ~13% over CI's 6.907) for the reason the 2026-08-05 note already gives: a cap pinned just above the measurement reds on the next commit and teaches the reader to raise it again. ALTERNATIVE EXAMINED AND REJECTED, so it is not re-proposed: dropping measure_turn_end_gate.ts from the pack would have cleared the cap with room, but the settings description added in the same change instructs CONSUMERS to run that script against their own transcript store before arming the gate — excluding it would break a shipped instruction.",
+            "baseline_note": "5.903 → 6.398 re-baselined 2026-08-05. This is a RE-BASELINE, not a carve-out for the change that tripped it: measured on the tracked tree, `origin/main` alone was already at 6.372 of a 6.4 cap, so 28 KB of headroom remained after months of accretion and the next legitimate addition of any size was going to cross it. The addition here is 26 KB — docs/contracts/settings-classes.md (the A/B/C fence, read at RUNTIME by the settings writer, so it must ship) plus src/server/io/yamlIO.ts (imported by that writer to preserve the user's comments). CI measures ~4 KB above the local figure because it regenerates dist/agent-src first. Max set to 6.9 rather than to a number that merely clears: a cap pinned 2 KB above the measurement reds on the next commit and teaches the reader to raise it again, which is how a budget becomes a formality. That restores the ~8% headroom the 5.903/6.4 pair originally carried. The `owner: maintainer` on this file stands — this raise is surfaced in the PR for that decision, not smuggled.",
+            "baseline_note_2026_08_24_lowered": "9.2 -> 9.1 LOWERED 2026-08-24 — the first DOWNWARD move in this file's history, and the reason it is possible is that road-to-npm-payload-reduction measured the lever the previous four raises kept deferring. Method, because the verdicts turn on it: a behavioural global-install harness (src/scripts/pack_install_smoke.ts) packs a real BUILT tarball, installs it into a throwaway prefix, and runs eleven consumer probes including the bash orchestrator FROM the installed tree, which is the only probe that exercises path-reachability rather than imports. Sensitivity was proven by sabotage before any verdict was trusted: excluding src/scripts/_lib/ produces ERR_MODULE_NOT_FOUND on four probes. WHAT THE BISECT FOUND, and it retires a suggestion this file has carried since 2026-08-20: EVERY named subtree SHIPS. src/scripts/ai_council/ — the 270 KB candidate the 2026-08-24 note left explicitly undecided — breaks council:status AND hooks:status with ERR_MODULE_NOT_FOUND. So does src/scripts/hooks/ (271.5 KB), src/scripts/_cli/ (204.1 KB), src/agent-src/ (101.8 KB), src/scripts/ai-video/ (95.4 KB, its shell entry point), src/scripts/mcp_server/ (53.1 KB). Two of those first read GREEN and flipped RED when the probe set was widened with mcp:setup and an ai-video shell check — recorded because it is the exact false-negative shape this file's own Risk 2 names, caught rather than shipped. WHAT IS ACTUALLY REMOVABLE is three FILE PATTERNS, not subtrees, and no prior note looked for them: dist/**/*.map (323.5 KB), dist/agent-src/skills/**/evals/** (56.2 KB), src/**/*.test.ts + *.spec.ts (22.6 KB) — 402.9 KB on the built tarball, all eleven probes green together. MEASURED under this file's own documented conditions (npm pack --dry-run --ignore-scripts, clean unbuilt worktree, which is what check_pack_size reads): 8.5596 -> 8.4667 MB, 2,775 -> 2,609 entries, a 92.9 KB drop. The gap between 402.9 KB and 92.9 KB is not a discrepancy: most of the .map saving lives in dist/cli, dist/ui, dist/mcp and dist/hooks, which --ignore-scripts excludes by design. HEADROOM IS 7.4 %, NOT 8 %, and the number is stated exactly rather than rounded up to match the pattern: 9.1 over the CI-adjusted 8.475 (local 8.4667 plus the ~8 KB CI delta every prior note here records) is 7.37 %. The 5.903/6.4, 7.238/7.8, 7.805/8.4 and 8.439/9.1 pairs carried ~8 %; this one carries slightly less, which is a real if small increase in flake risk and is why the falsifier below names it. STILL OPEN, quantified for the first time: src/agent-src/ is 94.2 % byte-identical to dist/agent-src/ (284 of 297 files, 2,144,870 of 2,276,707 bytes, content-hash diff). Two prior notes called this 'partial'; it is near-total. It is NOT removable by exclusion — nine shipped scripts under src/scripts/ import through it (build_archive_index, check_estate_count, eval_ui_triviality, explain_run, lint_ui_stack_bundles, skill_usage_report, smoke_quickstart, telemetry_disclosure_hook, utilization_report) while dist/agent-src/ is the tree the installer deploys, so the two copies ship for different reasons. Repointing those imports at dist/ would make source depend on a generated projection, which src/rules/source-of-truth.md exists to prevent. AI council 2026-08-24, 2/2: that needs an ADR covering compiled script entry points and asset resolution, not another files[] experiment. Worth ~102 KB and deliberately not taken here. TRADE-OFF ACCEPTED, named rather than absorbed: excluding dist/**/*.map means published packages carry no source maps, so consumer stack traces show compiled names. GREEN here means 'no regression across the recorded eleven-probe matrix', never 'semantically equivalent'. REVISIT-IF: three clean CI runs measure above 9.1 (the method does not match the gate); or observed run-to-run variance exceeds the 7.4 % margin, making the ratchet noise rather than a floor; or the src/agent-src/ ADR lands and the cap can drop again. NEVER raise it again from a fresh trip without the discriminator the 2026-08-10 note demands. DECIDED BY: AI council 2026-08-24, 2/2 convergent (anthropic/claude-sonnet-4-5 + openai/codex-default, 2 rounds, blind peer review), on a maintainer delegation of owner-reserved blockers for an autonomous drain run. Both seats independently required the 7.4 % to be stated exactly and the GREEN to be scoped to the probe matrix."
         }
     },
     "per_skill_share": {
@@ -26,13 +26,13 @@
             "design-intelligence": {
                 "max_pct": 23.5,
                 "measured_pct": 22.63,
-                "reason": "771 KB of BM25 corpus (27 CSVs + 16 design-language prose specs + manifest.json) consumed at runtime by corpus-grounding's retriever, and referenced by 6 further skills. Moving it behind a lazy fetch would add a network dependency to a skill that works offline today \u2014 the same trade-off ADR-061 \u00a78 already REJECTED for a 745 KB font index. The exception is visible debt, not an endorsement: raising max_pct requires a reason in the same commit.",
-                "baseline_note_2026_08_13": "23.0 \u2192 23.5 on 2026-08-13, and the reason belongs in this file per the sentence above. WHAT WAS ADOPTED: data/motion.csv, 10.5 KB, 16 rows. WHY IT IS NOT DISCRETIONARY: the roadmap's three design dials were all in its original scope, and two of the three (variance, density) were already portable \u2014 variance's keywords resolve against the local styles.csv (2/3/3 hits) and density is three pure range maps. The motion dial is not: upstream filters its corpus rows by `Intensity Tier == <dial tier>`, and that column exists only in motion.csv, so shipping the dial without the file means shipping a flag that silently does nothing. Two of three dials plus an inert third is a worse surface than three working ones. WHAT WAS DECLINED IN THE SAME PASS, so this is not a ratchet slipping: google-fonts.csv (743 KB \u2014 ADR-061 \u00a78 rejects it by name and by size; the 2026-06-16 amendment authorises a slim top-N slice from a different mirror, not this file), typography.csv (our curated font-pairings-reference.csv is the ADR's working substitution), six desktop-UI stacks outside frontend-design's scope, and eleven content-drifted files nobody has reviewed. The base pin in ATTRIBUTION.md is deliberately NOT bumped for the same reason. HONEST LIMIT ON THE NUMBER: 22.63 is the committed measurement, not a fresh one \u2014 check_pack_size cannot be re-measured from a built worktree (it reads 9.096 against the 7.8 packed cap before any change, the trap the packed_size_mb note above documents), so the post-adoption share is arithmetic (~22.87 %) and the new cap carries ~0.6 pp over it rather than clearing by a hair. DECIDED BY: AI council, 2026-08-13, on a maintainer delegation \u2014 and honestly, by ONE member over two rounds, not by convergence: the second seat returned exit_1 (the codex CLI refuses a git worktree as an untrusted directory). Recorded as a single-member review in agents/evidence/analysis/design-system-blockers-council.md."
+                "reason": "771 KB of BM25 corpus (27 CSVs + 16 design-language prose specs + manifest.json) consumed at runtime by corpus-grounding's retriever, and referenced by 6 further skills. Moving it behind a lazy fetch would add a network dependency to a skill that works offline today — the same trade-off ADR-061 §8 already REJECTED for a 745 KB font index. The exception is visible debt, not an endorsement: raising max_pct requires a reason in the same commit.",
+                "baseline_note_2026_08_13": "23.0 → 23.5 on 2026-08-13, and the reason belongs in this file per the sentence above. WHAT WAS ADOPTED: data/motion.csv, 10.5 KB, 16 rows. WHY IT IS NOT DISCRETIONARY: the roadmap's three design dials were all in its original scope, and two of the three (variance, density) were already portable — variance's keywords resolve against the local styles.csv (2/3/3 hits) and density is three pure range maps. The motion dial is not: upstream filters its corpus rows by `Intensity Tier == <dial tier>`, and that column exists only in motion.csv, so shipping the dial without the file means shipping a flag that silently does nothing. Two of three dials plus an inert third is a worse surface than three working ones. WHAT WAS DECLINED IN THE SAME PASS, so this is not a ratchet slipping: google-fonts.csv (743 KB — ADR-061 §8 rejects it by name and by size; the 2026-06-16 amendment authorises a slim top-N slice from a different mirror, not this file), typography.csv (our curated font-pairings-reference.csv is the ADR's working substitution), six desktop-UI stacks outside frontend-design's scope, and eleven content-drifted files nobody has reviewed. The base pin in ATTRIBUTION.md is deliberately NOT bumped for the same reason. HONEST LIMIT ON THE NUMBER: 22.63 is the committed measurement, not a fresh one — check_pack_size cannot be re-measured from a built worktree (it reads 9.096 against the 7.8 packed cap before any change, the trap the packed_size_mb note above documents), so the post-adoption share is arithmetic (~22.87 %) and the new cap carries ~0.6 pp over it rather than clearing by a hair. DECIDED BY: AI council, 2026-08-13, on a maintainer delegation — and honestly, by ONE member over two rounds, not by convergence: the second seat returned exit_1 (the codex CLI refuses a git worktree as an untrusted directory). Recorded as a single-member review in agents/evidence/analysis/design-system-blockers-council.md."
             }
         }
     },
     "built_surface_measurement_2026_08_24": {
-        "purpose": "road-to-standing-payload-truth 2.2 and 2.3. Every cap raise in this file's baseline_note_* history was measured on the UNBUILT tree, and nothing recorded what the BUILT surface \u2014 the one a consumer installs \u2014 actually is. Two figures for one lever with no label is how the confusion this file warns about starts, so both are recorded here with the command that produced each.",
+        "purpose": "road-to-standing-payload-truth 2.2 and 2.3. Every cap raise in this file's baseline_note_* history was measured on the UNBUILT tree, and nothing recorded what the BUILT surface — the one a consumer installs — actually is. Two figures for one lever with no label is how the confusion this file warns about starts, so both are recorded here with the command that produced each.",
         "unbuilt": {
             "packed_mb": 8.4953,
             "unpacked_mb": 28.7516,
@@ -45,7 +45,7 @@
             "unpacked_mb": 37.3775,
             "entries": 2808,
             "command": "npm run build && npm pack --dry-run --json   # prepack runs the build; this is what publish emits",
-            "note": "The surface a consumer downloads and installs. Gated by nothing \u2014 see _comment."
+            "note": "The surface a consumer downloads and installs. Gated by nothing — see _comment."
         },
         "delta": {
             "packed_mb": 2.0572,
@@ -53,8 +53,9 @@
             "entries": 194,
             "correction": "The roadmap and an earlier note in this file put the built artifact at '~5.4 MB larger'. MEASURED it is +8.6259 MB unpacked and +2.0572 MB packed. The 5.4 figure is stale and is corrected here rather than carried; a reader comparing 8.4953 against 28.22 or 37.3775 was comparing three different surfaces."
         },
-        "what_this_says_about_the_removed_gate": "unpacked_size_mb was removed on 2026-08-04 with max 28 and a measurement of 28.22 \u2014 0.8 % over. The built unpacked size TODAY is 37.3775 MB, i.e. that cap would now be exceeded by 33.5 %. Reinstating it is therefore not 'add a key with a measured baseline': it is setting a cap a third higher than the one whose 0.8 % breach caused its removal. Recorded because it makes the 2026-08-04 decision look better in hindsight than it did at the time, and because an owner reopening it under decision-revisit-gate should start from 37.4, not from 28.",
+        "what_this_says_about_the_removed_gate": "unpacked_size_mb was removed on 2026-08-04 with max 28 and a measurement of 28.22 — 0.8 % over. The built unpacked size TODAY is 37.3775 MB, i.e. that cap would now be exceeded by 33.5 %. Reinstating it is therefore not 'add a key with a measured baseline': it is setting a cap a third higher than the one whose 0.8 % breach caused its removal. Recorded because it makes the 2026-08-04 decision look better in hindsight than it did at the time, and because an owner reopening it under decision-revisit-gate should start from 37.4, not from 28.",
         "measured_at": "2026-08-24",
         "measured_on": "One machine, two worktrees of the same commit. CI has historically read ~8 KB above the local packed figure; no CI reading of the BUILT surface exists, so the built numbers here are local-only and are not a CI baseline."
-    }
+    },
+    "built_surface_enforcement_2026_08_30": "How the two surfaces this file records are now ENFORCED, which until 2026-08-30 they were not. `check_pack_size` compared every payload against `budgets.packed_size_mb.max`, and that cap was measured on the UNBUILT tree — this file has said so since 2026-08-24 in `measurement_conditions` and in `built_surface_measurement_2026_08_24`, and the gate did not read it. Measured on the day this note was written: main packed 8.985 MB over 2715 entries with `--ignore-scripts` and passed; a branch adding six source files packed 9.922 over 2827 and failed `max: 9.1`. The 112-entry difference is `dist/cli` plus `dist/cli-delegate` — the build, not the diff. The gate now reads the payload it is judging (`payloadIsBuilt`, keyed on `dist/cli/**`, already used by the content classes for the same reason) and routes: an UNBUILT payload against `max` and the `last_measured` creep line, unchanged; a BUILT payload against `built_surface_measurement_*.built.packed_mb` with the same `regression_pct`. A built payload with no recorded built figure is a FINDING, not a pass — otherwise removing the record would silently disable the built path. What this does NOT do: it does not raise `max`, and it does not gate the built surface against an absolute cap. The built surface gets a regression line only, because no absolute figure for it was ever set by a maintainer and inventing one here would be exactly the invented number the `_comment` warns about. Revisit-if: a maintainer sets an absolute built cap, or the built figure is re-measured (add a new `built_surface_measurement_<date>` key — the newest is read, so no code changes)."
 }

package/src/config/preamble-payload-budget.json

@@ -1,5 +1,5 @@
 {
-    "_comment": "Per-spawn preamble payload ceiling (road-to-cache-economy Phase 3), unblocked by C-3 measuring 38.0% against a 15% bar. This is a RATCHET, not the roadmap's literal candidate numbers: those (median 40k / p95 50k) were anchored to an upstream ~37k baseline, while this package's own deterministic in-repo payload is ~102.5k tokens \u2014 a hard 40k gate would be red on the day it lands and would teach the reader to ignore it. The ratchet fails on GROWTH instead, so the number can only walk down, and the 40k/50k target stays recorded as the destination. Only the deterministic, in-repo buckets are gated: user-scope rules depend on what the developer happens to have installed globally and are therefore not CI-checkable.",
+    "_comment": "Per-spawn preamble payload ceiling (road-to-cache-economy Phase 3), unblocked by C-3 measuring 38.0% against a 15% bar. This is a RATCHET, not the roadmap's literal candidate numbers: those (median 40k / p95 50k) were anchored to an upstream ~37k baseline, while this package's own deterministic in-repo payload is ~102.5k tokens — a hard 40k gate would be red on the day it lands and would teach the reader to ignore it. The ratchet fails on GROWTH instead, so the number can only walk down, and the 40k/50k target stays recorded as the destination. Only the deterministic, in-repo buckets are gated: user-scope rules depend on what the developer happens to have installed globally and are therefore not CI-checkable.",
     "schema_version": 1,
     "registered_at": "2026-07-30",
     "owner": "maintainer",
@@ -12,11 +12,11 @@
             "CLAUDE.md hierarchy (project + project-local)"
         ],
         "excluded_buckets": [
-            "user-scope rules (~/.claude/rules/*.md) \u2014 machine-dependent, not CI-checkable",
-            "tool definitions + dispatch prompt \u2014 no local source carries the request payload; the census reports it as an explicit residual",
-            "personas (src/agent-src/personas, 36 files, ~37,181 tok) \u2014 measured ZERO standing cost 2026-08-22: .claude/personas/ is written by this repo's generators (5 in-repo references) and read by no host contract; the host's real subagent surface is .claude/agents/, which this package writes 1 file into. Reversal condition: a host publishing a personas contract, or a generator folding persona prose into an instruction file that IS read. Evidence: agents/evidence/analysis/uncounted-prose-corpora-load-semantics.md",
-            "contexts (src/agent-src/contexts, 58 files, ~114,042 tok) \u2014 measured ON-DEMAND 2026-08-22: 0 rules declare load_context_eager: (17 declare the lazy load_context:), and rule.schema.json:49 documents eager as the variant that counts against the per-rule budget. This records that no eager declaration exists, NOT that eager loading would be free \u2014 nobody has declared one, so nobody has measured it. Evidence: agents/evidence/analysis/uncounted-prose-corpora-load-semantics.md",
-            "skill bodies (src/skills/*/SKILL.md, 290 files, ~595,843 tok) \u2014 measured ON-DEMAND 2026-08-22: censusSkillsCatalog (preamble_byte_census.ts:216-234) builds the gated catalog bucket from `- name: description` per skill and reads no body byte, which is why 290 skills totalling ~595,843 tok report as ~14.4k standing. Corollary: the description cap is the WHOLE skill-side lever, not a fragment of it. Evidence: agents/evidence/analysis/uncounted-prose-corpora-load-semantics.md"
+            "user-scope rules (~/.claude/rules/*.md) — ALREADY COUNTED, inside the `project-scope rules` gated bucket; excluded here only to avoid double-counting. CORRECTED 2026-08-30 (road-to-agent-turnaround 4.2, AI council 2/2 over two rounds). The previous reason read \"machine-dependent, not CI-checkable\", and that was FALSE in both halves: the layer is written by this package's own installer out of dist/agent-src/rules/, and which rules reach the GLOBAL layer rather than the project one is decided by frontmatter (`workspaces:` naming the maintainer workspace), not by anything on the developer's machine. `preamble_byte_census.censusDeliveredRulePayload` computes it from the projection alone, reading nothing under ~/. Measured 2026-08-30: 119 projected − 15 package-only = 104 delivered, matching the installed file count exactly; the 1.5 % char gap against the machine reading is install drift. Round 1 of the council voted to GATE the bucket on the premise that its bytes were invisible to CI; round 2 was given the checker's bucket definitions, found the 104 are a SUBSET of the 119 the gate already measures in full, and chose (a′): correct the reason, add the reconciliation test, leave the baseline alone. Adding the bucket would have moved the baseline ~111k for zero additional delivered payload. What remains genuinely unreported by this GATE is the 104 / 13 / 2 destination split, which the CENSUS now prints — the council placed it there until distinct destinations need independent budget enforcement.",
+            "tool definitions + dispatch prompt — no local source carries the request payload; the census reports it as an explicit residual",
+            "personas (src/agent-src/personas, 36 files, ~37,181 tok) — measured ZERO standing cost 2026-08-22: .claude/personas/ is written by this repo's generators (5 in-repo references) and read by no host contract; the host's real subagent surface is .claude/agents/, which this package writes 1 file into. Reversal condition: a host publishing a personas contract, or a generator folding persona prose into an instruction file that IS read. Evidence: agents/evidence/analysis/uncounted-prose-corpora-load-semantics.md",
+            "contexts (src/agent-src/contexts, 58 files, ~114,042 tok) — measured ON-DEMAND 2026-08-22: 0 rules declare load_context_eager: (17 declare the lazy load_context:), and rule.schema.json:49 documents eager as the variant that counts against the per-rule budget. This records that no eager declaration exists, NOT that eager loading would be free — nobody has declared one, so nobody has measured it. Evidence: agents/evidence/analysis/uncounted-prose-corpora-load-semantics.md",
+            "skill bodies (src/skills/*/SKILL.md, 290 files, ~595,843 tok) — measured ON-DEMAND 2026-08-22: censusSkillsCatalog (preamble_byte_census.ts:216-234) builds the gated catalog bucket from `- name: description` per skill and reads no body byte, which is why 290 skills totalling ~595,843 tok report as ~14.4k standing. Corollary: the description cap is the WHOLE skill-side lever, not a fragment of it. Evidence: agents/evidence/analysis/uncounted-prose-corpora-load-semantics.md"
         ],
         "_bucket_classification_note": "All three additions above are EXCLUSIONS, not gated buckets, and that is the finding rather than a shortcut. A census bucket over a corpus measured at zero is a permanently green gate certifying nothing. The exclusions carry their measurement so the next pass that proposes one of these corpora as a token lever meets the number instead of the silence."
     },
@@ -30,7 +30,7 @@
         {
             "tokens": 102520,
             "at": "2026-07-31",
-            "why": "free tightening: the measured tree came in 79 tok under the registered baseline, so the ratchet walks down to meet it. Not a claimed reduction \u2014 nothing was optimised; the point of a ratchet is that a lower measurement becomes the new ceiling instead of becoming unused headroom."
+            "why": "free tightening: the measured tree came in 79 tok under the registered baseline, so the ratchet walks down to meet it. Not a claimed reduction — nothing was optimised; the point of a ratchet is that a lower measurement becomes the new ceiling instead of becoming unused headroom."
         }
     ],
     "headroom_pct": 5,
@@ -71,7 +71,7 @@
         "measured_cold_start_median_tokens": 230930,
         "c3_scope_dedup_reduction_pct": 38,
         "c3_condition": "FIXTURE-ONLY as of 2026-07-31: the 38% was measured on a byte-identical two-scope fixture, and that condition is currently unreachable for every consumer because install.ts:2723/2725 stamp package:/source_path: into every installed rule while the in-repo projection stamps nothing. Aligning versions yields 0/110 twins, not 110/110. The figure is the mechanism's ceiling, NOT a realised saving, and this baseline therefore assumes no dedup. Reachability is the subject of road-to-dedup-reachability; run measure_scope_dedup for the live split.",
-        "_comment": "The measured median comes from Claude Code transcripts (deduped by message.id + requestId) and is context for the ratchet, never the gated number \u2014 it moves with the developer's own session history."
+        "_comment": "The measured median comes from Claude Code transcripts (deduped by message.id + requestId) and is context for the ratchet, never the gated number — it moves with the developer's own session history."
     },
     "ci_delivery": {
         "workflow": ".github/workflows/standing-payload-delta.yml",

package/src/config/profiles/balanced.ini

@@ -17,3 +17,30 @@ discipline_profile=auto
 
 ; Chat history (crash recovery) — balanced = phase-level traceability
 chat_history_frequency=per_phase
+
+; ── Intended standing payload (road-to-delivered-cost-truth 1.2) ─────────
+;
+; The DENOMINATOR of `payload_amplification` — what this profile intends to
+; deliver, NOT what it currently measures. A first cut declared the measurement
+; and made the ratio 1.00 by construction; a number that cannot move is not a
+; measurement. The gap between intended and delivered is the whole point.
+;
+; The figures are the published milestone in `src/config/preamble-payload-budget.json`
+; (`baseline_tokens: 102520`, due 2026-11-10), split by category in the measured
+; proportion — rules carry 90.2 % of the standing set and skill catalogue lines
+; 9.8 % (asset_delivery_ledger, exact BPE, 2026-08-28).
+;
+; All three shipped profiles declare the SAME figures, and that is the finding
+; rather than an oversight: the tier knob (`rule_loading_tier`) selects which
+; rules ACTIVATE, not which are written into the preamble, so no shipped profile
+; delivers a smaller standing payload than any other. Inventing three different
+; numbers would imply a difference this suite does not have.
+;
+; `config_cost report` VERIFIES this before using it: a declaration above what
+; the measured tree can deliver is refused, because a profile cannot intend to
+; deliver more than exists and an inflated denominator makes any delivery look
+; disciplined. Both AI council seats (2026-08-28, 2/2) raised that gaming
+; direction independently.
+declared_rules_tokens=92473
+declared_skill_catalogue_tokens=10047
+declared_payload_sanity_headroom=0.10

package/src/config/profiles/full.ini

@@ -15,3 +15,30 @@ discipline_profile=full
 
 ; Chat history (crash recovery) — full = tool-level traceability
 chat_history_frequency=per_tool
+
+; ── Intended standing payload (road-to-delivered-cost-truth 1.2) ─────────
+;
+; The DENOMINATOR of `payload_amplification` — what this profile intends to
+; deliver, NOT what it currently measures. A first cut declared the measurement
+; and made the ratio 1.00 by construction; a number that cannot move is not a
+; measurement. The gap between intended and delivered is the whole point.
+;
+; The figures are the published milestone in `src/config/preamble-payload-budget.json`
+; (`baseline_tokens: 102520`, due 2026-11-10), split by category in the measured
+; proportion — rules carry 90.2 % of the standing set and skill catalogue lines
+; 9.8 % (asset_delivery_ledger, exact BPE, 2026-08-28).
+;
+; All three shipped profiles declare the SAME figures, and that is the finding
+; rather than an oversight: the tier knob (`rule_loading_tier`) selects which
+; rules ACTIVATE, not which are written into the preamble, so no shipped profile
+; delivers a smaller standing payload than any other. Inventing three different
+; numbers would imply a difference this suite does not have.
+;
+; `config_cost report` VERIFIES this before using it: a declaration above what
+; the measured tree can deliver is refused, because a profile cannot intend to
+; deliver more than exists and an inflated denominator makes any delivery look
+; disciplined. Both AI council seats (2026-08-28, 2/2) raised that gaming
+; direction independently.
+declared_rules_tokens=92473
+declared_skill_catalogue_tokens=10047
+declared_payload_sanity_headroom=0.10

package/src/config/profiles/minimal.ini

@@ -15,3 +15,30 @@ discipline_profile=off
 
 ; Chat history (crash recovery) — minimal = lightest footprint
 chat_history_frequency=per_turn
+
+; ── Intended standing payload (road-to-delivered-cost-truth 1.2) ─────────
+;
+; The DENOMINATOR of `payload_amplification` — what this profile intends to
+; deliver, NOT what it currently measures. A first cut declared the measurement
+; and made the ratio 1.00 by construction; a number that cannot move is not a
+; measurement. The gap between intended and delivered is the whole point.
+;
+; The figures are the published milestone in `src/config/preamble-payload-budget.json`
+; (`baseline_tokens: 102520`, due 2026-11-10), split by category in the measured
+; proportion — rules carry 90.2 % of the standing set and skill catalogue lines
+; 9.8 % (asset_delivery_ledger, exact BPE, 2026-08-28).
+;
+; All three shipped profiles declare the SAME figures, and that is the finding
+; rather than an oversight: the tier knob (`rule_loading_tier`) selects which
+; rules ACTIVATE, not which are written into the preamble, so no shipped profile
+; delivers a smaller standing payload than any other. Inventing three different
+; numbers would imply a difference this suite does not have.
+;
+; `config_cost report` VERIFIES this before using it: a declaration above what
+; the measured tree can deliver is refused, because a profile cannot intend to
+; deliver more than exists and an inflated denominator makes any delivery look
+; disciplined. Both AI council seats (2026-08-28, 2/2) raised that gaming
+; direction independently.
+declared_rules_tokens=92473
+declared_skill_catalogue_tokens=10047
+declared_payload_sanity_headroom=0.10

package/src/config/publish-surface.json

@@ -1,5 +1,5 @@
 {
-  "_comment": "GENERATED by src/scripts/check_publish_surface.ts --write. Do not hand-edit. It records the resolved publish surface so a change to what ships is a diff a reviewer reads. `files[]` is an ALLOWLIST and it overrides .npmignore for anything under an included root — measured, after an attempt to strip compiled test artefacts via .npmignore had no effect because dist/ is an included root. That is why `negations` exists and why `npmignore_shadowed` is reported: a pattern the live payload contradicts is present and not doing what its author expected.",
+  "_comment": "GENERATED by check_publish_surface --write. Do not hand-edit. It records the resolved publish surface so a change to what ships is a diff a reviewer reads. `files[]` is an ALLOWLIST and it overrides .npmignore for anything under an included root — measured, after an attempt to strip compiled test artefacts via .npmignore had no effect because dist/ is an included root. That is why `negations` exists and why `npmignore_shadowed` is reported: a pattern the live payload contradicts is present and not doing what its author expected.",
   "roots": [
     "dist/agent-src/",
     ".augment-plugin/",
@@ -23,8 +23,11 @@
     "AGENTS.md",
     "CHANGELOG.md",
     "CONTRIBUTING.md",
+    "CREDITS.md",
     "LICENSE",
     "MIGRATION.md",
+    "NOTICE",
+    "docs/THIRD-PARTY-NOTICES.md",
     "README.md",
     "llms.txt"
   ],
@@ -32,7 +35,11 @@
     "!dist/**/*.test.js",
     "!dist/**/*.test.js.map",
     "!dist/**/*.spec.js",
-    "!dist/**/*.spec.js.map"
+    "!dist/**/*.spec.js.map",
+    "!src/**/*.test.ts",
+    "!src/**/*.spec.ts",
+    "!dist/agent-src/skills/**/evals/**",
+    "!dist/**/*.map"
   ],
   "npmignore": [
     ".idea/",

package/src/config/routing-coverage-seed.json

@@ -10,11 +10,11 @@
     },
     "seed": {
         "rules": 0.8952,
-        "skills": 0.301
+        "skills": 0.3344
     },
     "seed_provenance": {
         "rules": "94 covered of 105 routed rules (tier-1 21/24, tier-2 73/81) on 2026-08-25.",
-        "skills": "76 of 299 skills carry evals/triggers.json on 2026-08-25.",
+        "skills": "100 of 299 skills carry evals/triggers.json on 2026-08-30.",
         "finding": "The gap between the two seeds IS the roadmap defect D1, quantified. The rules surface is ~90% covered by a deterministic corpus that can fail a PR; the skills surface -- the one production routes on -- is 25%, covered only by a harness that is advisory only, never gating (rule_trigger_eval.ts:4)."
     },
     "history": [
@@ -24,6 +24,20 @@
             "from": 0.2542,
             "to": 0.301,
             "reason": "road-to-routing-assurance Phase 2.2 authored 14 high-tier corpus files (76 -> 90 of 299). The seed rises so the gain cannot be given back silently; it may never be LOWERED."
+        },
+        {
+            "date": "2026-08-26",
+            "scope": "skills",
+            "from": 0.301,
+            "to": 0.3144,
+            "reason": "road-to-canonical-terms 2.3 authored 4 corpus files while clearing check-trigger-eval-presence, which was red on main with 18 findings (90 -> 94 of 299). The seed rises for the same reason the previous entry did, and the reason is not bookkeeping: this file's own test asserts the seed IS the live measurement, so a gain left unrecorded reds the suite for doing exactly what the ratchet exists to encourage. It may never be LOWERED."
+        },
+        {
+            "date": "2026-08-30",
+            "scope": "skills",
+            "from": 0.3144,
+            "to": 0.3344,
+            "reason": "road-to-governed-harness-evolution Phase 2.4 authored six wave-1 corpus files (94 -> 100 of 299), chosen by 2.2's inbound-confusion-edge ordering rather than picked. The seed rises so the gain cannot be given back silently; it may never be LOWERED. Surfaced by check_routing_coverage.test.ts, which asserts seed == live measurement in BOTH directions -- a seed left below the tree is a red test, not slack, which is the same asymmetry check_source_size_budget uses and for the same reason: nothing else would force the ratchet to follow a gain."
         }
     ]
 }