experimental-design
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves ingesting and processing external datasets and local files, which presents a surface for indirect prompt injection.\n
- Ingestion points: The workflow involves inspecting local files and fetching external datasets via the
sandbase_runtool as described inSKILL.md.\n - Boundary markers: The instructions lack specific guidance on using delimiters or safety markers when the agent processes these external data sources.\n
- Capability inventory: The skill is designed to perform local analysis and implementation, implying file access and execution capabilities, alongside network access to the SandBase API.\n
- Sanitization: There are no instructions for sanitizing or filtering input from external data sources before it is used in transformations or analysis.
Audit Metadata