dsar-privacy
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing external requests which represents a theoretical injection surface.
- Ingestion points: Untrusted requester names, identifiers, and request details are ingested into template fields across Stage 1, Stage 3, and Stage 5.
- Boundary markers: Absent. The templates use simple placeholders like
[name]or[Requester]without explicit instructions to the agent to treat this content as data only or to ignore embedded instructions. - Capability inventory: None. The skill consists entirely of markdown instructions and does not include scripts, command execution tools, or network capabilities.
- Sanitization: Absent. No instructions are provided for the agent to validate or sanitize external input before inclusion in outputs.
- [NO_CODE]: The skill is entirely instructional and does not include any executable scripts, binary files, or external dependencies. It relies solely on natural language instructions for the agent.
- [SAFE]: The skill enforces multiple security and compliance best practices, including mandatory escalation to legal counsel for sensitive actions (Erasure, Portability, Objection) and a strict requirement for licensed attorney review of all outputs before they are sent to external parties.
Audit Metadata