cro
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze marketing pages, landing pages, and forms based on URLs provided by the user. This creates an attack surface where malicious instructions embedded in the analyzed web pages could potentially influence the agent's output or actions.
- Ingestion points: The skill specifically prompts the user for URLs to analyze as part of its core functionality, as seen in the
SKILL.mddescription and the test cases inevals/evals.json. - Boundary markers: The instructions do not define clear boundaries, delimiters, or explicit "ignore" directives for the agent when processing content retrieved from external URLs.
- Capability inventory: The skill is strictly instructional and contains no executable scripts. It lacks high-risk capabilities like arbitrary command execution, network exfiltration, or file system modification.
- Sanitization: There is no evidence of sanitization, filtering, or validation of the external content before it is incorporated into the agent's analysis context.
Audit Metadata