gpai-code-of-practice
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted user-supplied data regarding AI model capabilities, training compute, and data provenance to produce compliance assessments.
- Ingestion points: Untrusted data enters the agent context via the 'Quick Question Set' in SKILL.md and user responses throughout the 6-step workflow.
- Boundary markers: The skill does not employ explicit delimiters or XML tags to isolate user data from the agent's instructions, relying on natural language context.
- Capability inventory: The skill is limited to generating text-based deliverables (Gap Assessment, GPAI Compliance Memo, and Executive Summary) and does not invoke external tools, file system writes, or network operations.
- Sanitization: There is no evidence of input validation or sanitization of the user-provided data before it is interpolated into the provided output templates.
Audit Metadata