08-operations-plan
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied data to generate business plans, creating a theoretical attack surface for indirect prompt injection.
- Ingestion points: Inputs such as "Offer definition", "Supplier information", and "Technology infrastructure" provided by the user in the prompt are used to populate the Section 08 Operations Plan.
- Boundary markers: The
SKILL.mdfile contains a "Capability and Permission Boundaries" section that explicitly restricts the skill to "read-only analysis and drafting," providing a clear instruction to the agent regarding its operational limits. - Capability inventory: No scripts or instructions within the skill utilize subprocess calls, network operations, file writing, or dynamic code execution.
- Sanitization: The skill uses structured templates (e.g., the "Process Documentation Format") for output, which helps maintain the context, but does not explicitly implement input sanitization for user-provided data.
Audit Metadata