scholar-sidekick-python
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
scholar-sidekickPython package from PyPI. This is the official client library for the service described in the skill documentation.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted citation data, creating a potential surface for indirect prompt injection attacks.\n - Ingestion points: Untrusted citation metadata (titles, authors) and raw reference strings are ingested via the
verify(),audit(),audit_bibliography(), andformat()methods inSKILL.md.\n - Boundary markers: The documentation does not show the use of explicit delimiters or specific instructions to the agent to ignore instructions embedded within citation data.\n
- Capability inventory: The skill is capable of performing network operations by making requests to the
scholar-sidekick.comREST API.\n - Sanitization: There is no evidence in the provided scripts or documentation of input validation or sanitization for the citation data handled by the client.
Audit Metadata