1k-browser-module
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a large surface area for processing untrusted data, specifically DApp requests and external web content.
- Ingestion points: Data enters the system through the JSBridge, remote search API results, and URLs provided to the WebView lifecycle hooks.
- Boundary markers: The instructions consistently mandate the use of uriUtils.validateUrl, validateWebviewSrc, and parseDappRedirect to inspect and block malicious or unsupported content before processing.
- Capability inventory: The module has the capability to open transactional modals, persist data to the application's SimpleDB, and interact with browser tabs.
- Sanitization: The Hard Rules explicitly prohibit passing secrets or unbounded payloads through modal parameters or loggers, ensuring sensitive information is kept out of potentially exposed channels.
- [EXTERNAL_DOWNLOADS]: The documentation references several API endpoints for discovery services (e.g., /utility/v1/discover/dapp/homepage). As these represent the vendor's (onekeyhq) own backend services required for the DApp discovery functionality, they are documented as legitimate resources within the skill's scope.
Audit Metadata