tw-research-citation-checker

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to call and parse third-party search/results (Step B/C: Consensus MCP search and WebSearch including Google Scholar, CrossRef/doi.org, CNKI, etc.) and to use those results to decide existence and next actions, so it ingests public third‑party content that can materially influence behavior.