landing-page-design
Pass
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No instructions targeting the bypass of safety filters or modification of system behavior were found. The prompt templates use standard placeholders for user-specified design topics.
- [Data Exposure & Exfiltration] (SAFE): The file contains no hardcoded credentials, sensitive file path references, or network operations that could exfiltrate data.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): No package managers, remote script downloads, or dynamic execution patterns (like eval or subprocess) are present.
- [Indirect Prompt Injection] (LOW): The skill identifies user input as a core part of its design templates. 1. Ingestion points: User inputs for product types and emotions in the prompt_templates section. 2. Boundary markers: Explicit delimiters for user input are absent. 3. Capability inventory: The skill has no active tools or code execution capabilities. 4. Sanitization: No sanitization or validation of the ingested strings is performed.
Audit Metadata