ux-tenets-traps
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process evidence from external, potentially untrusted sources such as source code, Figma frames, and screenshots. This creates a surface where maliciously crafted content within those artifacts could attempt to influence the agent's output or subsequent actions.\n
- Ingestion points: The workflow in
SKILL.mdandreferences/screenshot-audit.mdexplicitly gathers evidence from code, screenshots, and Studio artifacts.\n - Boundary markers: The skill lacks explicit delimiters or instructions to the agent to disregard instructions embedded within the analyzed artifacts.\n
- Capability inventory: The skill utilizes
memiCLI tools for auditing and diagnosis based on the gathered evidence.\n - Sanitization: No sanitization, escaping, or validation of the input artifact content is specified before the data is processed.
Audit Metadata