cardiology-science-for-people
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources via PubMed MCP tools (search_articles, get_article_metadata, get_full_text_article), creating a surface for indirect prompt injection. Maliciously crafted content within indexed scientific abstracts or metadata could theoretically influence the agent, as the instructions lack explicit boundary markers or sanitization requirements for retrieved data.\n
- Ingestion points: The skill processes external trials, research papers, and user-provided PDFs via the PubMed MCP integration mentioned in the Workflow section of SKILL.md.\n
- Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined to isolate untrusted external content from the agent's core instructions.\n
- Capability inventory: The skill's capabilities are focused on text generation and research. No dangerous shell commands, file system writes, or unauthorized network exfiltration patterns were identified in the analysis.\n
- Sanitization: There are no explicit instructions for sanitizing, filtering, or escaping retrieved content before it is interpolated into the agent's context.
Audit Metadata