literature-review-agent
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches academic paper metadata from well-known technology and research services including Semantic Scholar (api.semanticscholar.org), Crossref (api.crossref.org), and OpenAlex (api.openalex.org).
- [EXTERNAL_DOWNLOADS]: Supports optional candidate discovery through third-party search providers Exa (api.exa.ai) and Tavily (api.tavily.com) using environment-stored API keys.
- [COMMAND_EXECUTION]: Orchestrates the pipeline by executing multiple bundled Python utility scripts (e.g., bibtex_format.py, cross_verify.py, s2_search.py) to process research data and manage API interactions.
- [INDIRECT_PROMPT_INJECTION]: The skill's ingestion of web search results creates a potential surface for indirect prompt injection. This is explicitly mitigated by the implementation of rigorous verification rules, including Levenshtein title matching and cross-index corroboration, which serve to validate the existence and metadata of all citations before they are integrated into the final output.
Audit Metadata