alignment
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from running web pages, which are untrusted external sources. This content can potentially contain malicious instructions aimed at the agent.\n
- Ingestion points: The script
measure.jsextracts text from elements usingtextContentand includes 40-character samples in the finding descriptions.\n - Boundary markers: The findings report does not utilize explicit boundary markers or instructions to treat extracted page text as untrusted data.\n
- Capability inventory: The agent can capture screenshots and is instructed to perform code modifications to fix alignment errors.\n
- Sanitization: The extracted text is truncated and trimmed but not specifically sanitized to remove or neutralize potential prompt injection attacks.
Audit Metadata