systematic-review
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources such as PubMed, arXiv, Semantic Scholar, and CrossRef. This creates a surface for indirect prompt injection attacks where malicious instructions could be embedded in paper titles, abstracts, or metadata to influence the agent's behavior during screening, extraction, or synthesis.
- Ingestion points: Bibliographic data and full-text PDFs from external databases (Step 2 and Step 4).
- Boundary markers: None explicitly defined in this meta-skill file.
- Capability inventory: Mentions integration with other skills for literature search, PDF extraction, and report generation.
- Sanitization: No specific sanitization or validation of the external content is mentioned in the workflow description.
- [NO_CODE]: This skill primarily consists of descriptive markdown instructions and workflow orchestration logic without any embedded executable scripts or code.
Audit Metadata