scholar-search
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and displays untrusted data from the open web and academic repositories via the AIsa API.
- Ingestion points: External data enters the agent context through the search results, snippets, and raw content extracted by
scripts/search_client.py. - Boundary markers: The script uses text separators (e.g.,
====================) to format output, but lacks explicit boundary markers or instructions directing the agent to ignore potential commands embedded in the results. - Capability inventory: The skill utilizes network communication via
urllib.requestand subprocess execution of the search script. - Sanitization: No validation or filtering is performed on titles, publication info, or abstracts retrieved from external sources before they are passed to the agent.
- [COMMAND_EXECUTION]: The skill instructions and examples direct the agent to execute
scripts/search_client.pyas a local subprocess to perform search operations.
Audit Metadata