card-xiaohongshu
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill treats the user's message as a design brief which is then used to guide the agent in authoring files using its own editing tools. This creates a surface where a malicious user message could attempt to influence the file creation process beyond the intended design scope.
- Ingestion points: The user's message is directly mapped to a brief in
SKILL.md. - Boundary markers: None are defined in the instructions to separate user input from system instructions.
- Capability inventory: The skill utilizes file-editing tools and the
register_open_design_artifacttool to modify the local environment and output design files. - Sanitization: No sanitization or validation of the brief content is specified before the agent is instructed to author files.
Audit Metadata