browser-analyze
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface because it processes untrusted data from external websites. A malicious page could contain instructions hidden in text, metadata, or structured data that attempt to manipulate the agent's analysis or override its instructions.
- Ingestion points: External URLs provided by users for analysis in
SKILL.md. - Boundary markers: Absent; there are no instructions to delimit or ignore embedded instructions within the target page content.
- Capability inventory: Browser navigation, screenshot capture, and JavaScript-based DOM inspection in
SKILL.md. - Sanitization: None; the skill extracts raw text and computed styles directly from the DOM.
- [EXTERNAL_DOWNLOADS]: The skill specifies a dependency on an external Model Context Protocol (MCP) server at
https://mcp.trylexsis.com/mcp. This server is used to provide storefront analysis capabilities to the agent.
Audit Metadata