web-automation
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected.\n
- Ingestion points: The skill ingests untrusted data from external web pages and DOM structures during its analyze and data extraction phases (as described in SKILL.md).\n
- Boundary markers: No explicit delimiters or directives are specified to instruct the agent to ignore instructions embedded within the processed web content.\n
- Capability inventory: The skill possesses the ability to click buttons, fill forms, and navigate between pages, which provides a functional path for instructions injected from external sites (as described in SKILL.md).\n
- Sanitization: While features like 'data anonymization' and 'safe operation with validation' are mentioned, there is no technical evidence of sanitization to prevent the agent from obeying instructions found on web pages.
Audit Metadata