config-audit/tests/lib
Kjell Tore Guttormsen e9921d3c9d feat(optimize): add --subtract, the subtraction axis, behind a deterministic floor
Every command so far asked an addition question — what to add, what to move,
what it costs. Nothing asked what is no longer earning its always-loaded rent.
This adds that axis as a fourth lensCheck on the existing hybrid motor rather
than a new scanner or a 22nd command: the measured payoff (~18% of one file)
justifies a mode, not machinery.

It is the only lens that proposes REMOVING config, so it carries a guarantee
the others don't need: a load-bearing block is never a candidate. Precision is
asymmetric — a missed dead line costs a few tokens per turn, a deleted one
costs a wrong remote or a broken script — so the floor is decided in code
(lib/floor-exclusion.mjs) before the opus judge sees anything, never in prose.

Granularity is the leaf block, with two structural exceptions: a paragraph
ending in ':' merges with the list it introduces, and an ordered list is a
contract whose steps inherit floor from any sibling. Unordered lists
deliberately do not inherit — a load-bearing bullet and a disposable one
routinely share a list, and container-reasoning is the error the hand-built
ground truth exists to catch.

Verified against that ground truth (built before any classifier existed), with
the comparison machine-checked rather than read by eye: zero load-bearing
blocks proposed, 11/18 deletable groups surfaced, ~756 tok ~ 18% of a ~4300
token file — inside the pre-registered band. The first run found five floor
violations the synthesized fixture missed; each got a structural rule and a
fixture shape so it cannot regress.

Three real bugs the dogfood run exposed, all now covered:
- JS \b is ASCII-only, so /\bunngå\b/ never matches — every Norwegian keyword
  ending in æ/ø/å was silently dead.
- A bare word/word is not a path; "pros/cons" vetoed the largest deletable
  block until PATH_RE was tightened to rooted paths and globs.
- "Mid-sentence" must key on a preceding lowercase letter; the loose version
  read **bold labels:** and quoted openers as entities, costing 4 of 11 groups.

BP-SUB-001 is grounded entirely in the Anthropic steering blog already cited by
BP-MECH-001..004 and asserts nothing from the talk that motivated the feature —
no "80%", no ablation figure.

Suite 1365 -> 1382/0. Frozen v5.0.0 snapshots untouched; plain optimize output
byte-identical on identical input (--subtract adds keys only when passed).
knowledge-refresh-cli's reference date moved to 2026-08-01: its premise that
every seed entry was verified 2026-06-20 expired when BP-SUB-001 got a genuine
verification date, and backdating the entry to fit the test would have been a
lie about when its source was checked.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RW2haJXbxZpKivKHseSXNh
2026-07-31 16:24:51 +02:00
..
active-config-reader.test.mjs fix(acr): token estimator discounts block-level HTML comments (M-BUG-6) 2026-06-26 14:29:24 +02:00
active-model.test.mjs feat(skl,cml): --context-window auto model→window probe (v5.12 B8b) [skip-docs] 2026-06-23 22:29:12 +02:00
baseline.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
best-practices-register.test.mjs feat(knowledge): best-practices register foundation — v5.7 Fase 1 Chunk 1 2026-06-20 22:37:47 +02:00
campaign-export.test.mjs feat(campaign): plan export + execution-by-reuse (v5.7 Fase 2 Block 4c) 2026-06-23 10:08:04 +02:00
campaign-ledger.test.mjs feat(campaign): machine-wide always-loaded token roll-up — pure layer (v5.9 B2a) 2026-06-23 16:57:31 +02:00
context-window.test.mjs feat(skl,cml): --context-window auto model→window probe (v5.12 B8b) [skip-docs] 2026-06-23 22:29:12 +02:00
diff-engine.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
file-discovery.test.mjs fix(acr): file-discovery skips backups/ dirs — never live config (M-BUG-8) 2026-06-26 12:23:43 +02:00
forbidden-words-data.test.mjs feat(humanizer): forbidden-words data file (tier1/2/3) 2026-05-01 16:53:37 +02:00
humanizer-data.test.mjs fix(acr): on-demand copy for oversized skill-body finding (M-BUG-16) 2026-06-30 13:30:22 +02:00
humanizer.test.mjs fix(acr): AGT findings humanize to "Wasted tokens" not "Other" (M-BUG-17) 2026-06-30 13:51:55 +02:00
knowledge-refresh.test.mjs feat(knowledge): knowledge-refresh — living register, deterministic stale core + web poll (v5.7 Chunk 3) 2026-06-21 19:22:57 +02:00
lens-prefilter.test.mjs feat(opt): optimization lens Chunk 2b — opus prose-judgment analyzer + /config-audit optimize 2026-06-21 14:01:18 +02:00
output.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
permission-rules.test.mjs feat(dis): flag forbidden-param permission rules CC silently ignores 2026-06-19 14:04:25 +02:00
report-generator.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
scoring-humanizer.test.mjs feat(humanizer): wire humanizer into posture and scoring scorecard 2026-05-01 17:38:03 +02:00
scoring.test.mjs feat(config-audit): severity-weighted scoreByArea (v5 F3) 2026-05-01 06:20:08 +02:00
severity.test.mjs feat(config-audit): export WEIGHTS from severity.mjs (v5 F3 prep) 2026-05-01 06:16:28 +02:00
skill-listing-budget.test.mjs feat(skl,cml): --context-window calibration, advisory when unknown (v5.11 B8) [skip-docs] 2026-06-23 21:44:52 +02:00
string-utils.test.mjs fix(acr): SET typo-gates unknown-key false positives (M-BUG-10) 2026-06-26 15:15:14 +02:00
subtraction-prefilter.test.mjs feat(optimize): add --subtract, the subtraction axis, behind a deterministic floor 2026-07-31 16:24:51 +02:00
suppression.test.mjs feat(ultraplan-local): v1.6.0 — /ultraresearch-local deep research command 2026-04-08 08:58:35 +02:00
yaml-parser.test.mjs feat(acr,yaml): v5.6 Foundation — load-pattern enumeration + block-seq parser 2026-06-20 16:57:52 +02:00