match
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting data from untrusted external sources, including live web pages, third-party applications, and design files, to extract technical metadata such as styles, tokens, and states.
- Ingestion points: Instructions in
SKILL.md(Section 1) direct the agent to open "the live page, the installed app, the design file or the screenshot." - Boundary markers: The skill does not define specific delimiters or instructions to isolate external data from the agent's internal prompt instructions.
- Capability inventory: The skill instructions assume the agent has tools to inspect live pages and apps, though no specific tools are restricted or granted in the YAML frontmatter.
- Sanitization: There are no instructions for sanitizing or filtering potentially malicious instructions that might be embedded in the HTML, metadata, or design files of the source content.
Audit Metadata