wos-parse-results
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from Web of Science (WoS) API responses and browser DOM content.
- Ingestion points: Data is ingested via the
textparameter in Mode A and thedocumentobject in Mode B. - Boundary markers: The skill does not implement specific delimiters or warnings to the agent regarding potential instructions embedded within the search results.
- Capability inventory: No dangerous capabilities such as file system access, network operations, or subprocess execution were detected within this specific skill.
- Sanitization: Basic sanitization is performed by stripping HTML tags from abstracts, but no validation is performed against potential prompt injection patterns in titles, authors, or other text fields.
Audit Metadata