color-palette-extractor
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from website URLs and images, which could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: In
SKILL.md, the workflow explicitly identifies "website URL" and "image file" as input sources for fetching and parsing CSS or pixel data. - Boundary markers: There are no instructions to use delimiters or specific prompt engineering to ignore embedded instructions within the fetched CSS comments or image metadata.
- Capability inventory: The skill instructs the agent to "fetch and parse the CSS," which implies the use of network-capable tools and parsing logic on external content.
- Sanitization: The skill lacks instructions for escaping or validating the content retrieved from external websites before processing it within the agent's context.
Audit Metadata