parallel-web-search

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md requires running parallel-cli search to fetch web search results (saved to /tmp/*.json) and instructs the agent to read and cite those open web excerpts as facts, so it clearly ingests untrusted public third-party content from the open web.