update-golden-values

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The runtime path download_golden_values.py pulls GitHub Actions artifacts from the specified workflow run ID (public/outsider-authored content) and overwrites tests/functional_tests/test_cases/**/golden_values_*.json, which are then read by compare_golden_values_kl.py and converted into LLM-visible prose/summary text—enabling indirect prompt injection via artifact contents.