The skill no longer modifies code. It emits exactly two artifacts under simplify/:
findings.json (machine-actionable ledger) + a ranked report. Every finding now REQUIRES
title/action/verify (locked in ledger.schema.json) so each is independently executable by
the calling agent. Removed: apply/report modes, Track A autofix execution, baseline gate,
verify-each-change loop, the include-tests switch (tests excluded by default). Reframed
SKILL.md, README, PATTERNS (deterministic/judgment), DETECTION, OPTIMIZATION (judge-only
model tiering), PORTABILITY, SVELTE, CONSOLIDATION accordingly.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Test files carry intentional literals/structure, so they were the dominant
remaining magic-number noise source on real projects. Step 1 now filters common
test globs (per-language) from the candidate set by default; the user can opt in
with "include tests / --include-tests". Recorded in ledger config.include_tests.
On software-house: 73/283 TS files are tests; findings 235 -> 180 by default.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Multi-agent review found the ledger's suggested_pattern keys could miss the
catalog: add svelte-snippets/svelte-diagnostic sections, declare the Python
identity-compare/membership/truthiness keys explicitly, and note that
extract-method/remove-dead-code are detector-less (no manifest entry).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Giancarmine SalucciandClaude Opus 4.8