b2c-custom-caches
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents patterns for ingesting data from external APIs and filesystem sources to populate custom cache regions, which constitutes a potential injection surface.
- Ingestion points: The skill utilizes
LocalServiceRegistryfor fetching external API responses anddw.io.FileReaderfor reading configuration files from theFile.IMPEXdirectory (SKILL.md). - Boundary markers: Code examples for loader functions do not demonstrate the use of delimiters or specific instructions to disregard directives embedded within the cached data.
- Capability inventory: The skill patterns include network connectivity via platform services and local filesystem read access.
- Sanitization: While the skill demonstrates standard
JSON.parseimplementation for processing external content, it does not show secondary validation or schema enforcement for the retrieved data structures.
Audit Metadata