claygent

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from the web, creating an attack surface for indirect prompt injection where malicious content on a external website could attempt to influence the agent's behavior.
  • Ingestion points: The skill instructs the agent to visit company domains and perform Google searches to extract data (SKILL.md).
  • Boundary markers: The prompt templates include instructions such as "Only use information found on the website" to provide a contextual boundary.
  • Capability inventory: The skill is purely instructional and does not include any executable scripts or tools with file system or command execution access.
  • Sanitization: There is no evidence of programmatic sanitization or input filtering for the data retrieved from external sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 08:51 AM
Security Audit — agent-trust-hub — claygent