ai-feedback-design-principles
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied content (feedback scenarios and existing designs) which are interpolated into the prompt. This creates a surface for indirect prompt injection.
- Ingestion points: Untrusted data enters the agent context via the
feedback_scenarioandcurrent_feedback_designfields inSKILL.md. - Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands within the input fields.
- Capability inventory: The skill possesses no tools, network access, file system writes, or execution capabilities.
- Sanitization: No sanitization or validation is applied to the input data.
- Risk Assessment: The lack of system capabilities mitigates the risk; a malicious input could only influence the text of the generated feedback analysis.
- [SAFE]: The skill consists solely of instructions and metadata for pedagogical evaluation. No patterns for credential exfiltration, remote code execution, persistence, or obfuscation were detected.
Audit Metadata