fix: llm_judge test category has no execution path in cogworks-eval.py

[williamhallatt]

test-case-template.jsonl and deployment-skill/test-cases.jsonl define llm_judge category cases. cogworks-eval.py only handles activation-based behavioral tests — no code path to run LLM-judge quality checks exists. Fix: annotate templates as aspirational or implement a runner in cogworks-eval.py. Part of tests/ audit #23-#32.

[williamhallatt]

Fixed in recent commits. See commit log on main.