table
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFEPROMPT_INJECTIONREMOTE_CODE_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It takes data from the user context and renders it into a generated HTML file using innerHTML without sanitization or escaping. This allows malicious data to execute JavaScript in the browser environment. Ingestion points: Data extracted from user context as described in SKILL.md. Boundary markers: None present. Capability inventory: Write tool and mcp__chrome-devtools for browser navigation. Sanitization: No escaping is performed in the renderCell function in TEMPLATE.md.
- [REMOTE_CODE_EXECUTION]: The skill performs dynamic code execution by writing a self-contained HTML and JavaScript file to the local system and navigating a browser to it. This mechanism is the primary function of the skill but lacks safeguards when handling untrusted data.
Audit Metadata