blink-storage
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill incorporates tools like
blink_storage_listthat retrieve file metadata from external storage, which represents a potential surface for indirect prompt injection if filenames are used to influence agent behavior. - Ingestion points: File metadata returned by
blink_storage_list(SKILL.md). - Boundary markers: Absent in the tool definitions.
- Capability inventory: File listing, deletion, and URL generation.
- Sanitization: No explicit sanitization of storage metadata is mentioned.
Audit Metadata