bloat(#236): delete orphan scripts/audit-subagent-models.sh

[BaseInfinity]

Summary

ROADMAP #236 (functional-bloat hunt) — orphan deletion via the spec'd git rm + run-tests method.

scripts/audit-subagent-models.sh — 85 lines, 0 references anywhere in the repo (verified across *.sh, *.yml, *.json, *.md, *.py). Sat ~3 months on Back Burner with skeptical gate ("Research if users care before shipping"). Not in package.json files: list, so never shipped to npm consumers. Pure repo dead weight.

(Replaces #317, which got stranded after origin/main was force-updated mid-flow.)

How discovered

for s in scripts/*.sh scripts/*.py; do
    name=$(basename "$s")
    refs=$(grep -rl "$name" --include="*.sh" --include="*.yml" --include="*.json" --include="*.md" --include="*.py" . | grep -v "/.git/\|node_modules\|__pycache__" | wc -l)
    echo "$refs refs: $name"
done | sort -n

Output: 6 scripts inventoried, 1 orphan (audit-subagent-models.sh at 0 refs; the other 5 had 3-6 refs each).

Test plan

All test suites green after git rm:

• tests/test-cli.sh — 78/78 PASS
• tests/test-plugin.sh — 25/25 PASS
• tests/test-doc-consistency.sh — 35/35 PASS
• tests/test-workflow-triggers.sh — 169/169 PASS
• tests/test-hooks.sh — 154/154 PASS
• tests/test-self-update.sh — 153/153 PASS

Zero failures. Confirms the 0-reference grep was authoritative.

ROADMAP update

#236 marked phase-partial DONE:

• v1.69.0 BASELINE-once-per-session (~12K tokens/heavy session saved)
• v1.70.0 TDD CHECK-once-per-session (~0.5-1.5K tokens saved)
• v1.71.0 SDLC SKILL.md Cross-Model Review trim (4995 → 4568 tokens)
• This PR: orphan script deleted

Categories not exercised this round (tests redundancy, CLI templates, ROADMAP archive-below-fold) re-trigger on maintainer call-out.

Back Burner "Subagent Model Compliance Audit" entry updated with note pointing at this commit; if demand resurfaces, restore from git history.

Reversibility

Single git revert. Script reachable forever via the deletion commit.