defensive-boundaries
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists entirely of instructional markdown documentation intended to guide the agent in implementing defensive boundaries and validating external inputs.
- [NO_CODE]: No scripts, executables, or command-line instructions are present in the skill files, precluding direct execution risks.
- [INDIRECT_PROMPT_INJECTION]: While the skill discusses handling untrusted data, it does not provide an automated ingestion surface or tool-chaining capability that would introduce an indirect injection risk to the agent itself.
Audit Metadata