dbs-theory-grounding
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied propositions and cases for theoretical analysis.\n
- Ingestion points: Input text provided in Step 1 for analysis.\n
- Boundary markers: Relies on task-based structure; no explicit delimiters for untrusted input data are specified.\n
- Capability inventory: Includes the ability to use network retrieval (curl, web search) for verifying academic sources.\n
- Sanitization: No specific sanitization mentioned; the process focuses on analytical decomposition of the input.\n- [EXTERNAL_DOWNLOADS]: The skill allows for the use of external tools like
curlto verify citations and theory details.\n - Context: This capability is secondary and is gated by strict principles that favor offline knowledge and respect user-defined networking prohibitions.
Audit Metadata