search-first

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md’s required “Full Mode (agent)” workflow instructs launching a researcher subagent that searches npm/PyPI, MCP servers, and GitHub/Web, which at runtime can fetch and ingest outsider-authored free text (e.g., package docs/readmes, issue/discussion content, web pages) into the LLM context for comparison and recommendation.