paper-autoraters

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflow explicitly requires resolving paper references via Semantic Scholar ("S2 verification" in SKILL.md and scripts/compute_f1.py) and looking up venue average citation counts from public sources (references/litreview-quality-prompt.md), so it fetches and ingests open/public third-party content that can materially affect scoring.