awwwards-sites
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze user-provided reference evidence (websites or design materials) to establish art direction. This creates a potential surface for indirect prompt injection if the reference materials contain malicious instructions hidden in metadata or content.
- Ingestion points: Reference evidence provided by the user (Step 1 in SKILL.md).
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within reference materials.
- Capability inventory: The agent's capabilities when using this skill are limited to generating web development code (HTML, CSS, JavaScript). It lacks permissions for sensitive file access or unauthorized network operations.
- Sanitization: There are no explicit requirements for sanitizing or validating the content of the reference evidence.
Audit Metadata