hyperbrowser-mcp

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill configuration instructs the agent to use npx -y hyperbrowser-mcp, which downloads and executes the Hyperbrowser MCP server package from the official npm registry.
  • [INDIRECT_PROMPT_INJECTION]: The skill provides tools such as crawl_page and extract_data that ingest and process data from external, untrusted websites.
  • Ingestion points: Tools retrieve rendered DOM contents, link trees, and visual elements from arbitrary URLs provided during execution (SKILL.md).
  • Boundary markers: There are no explicit instructions or delimiters defined to isolate the retrieved web content or warn the model to ignore embedded instructions within the crawled data.
  • Capability inventory: The agent can perform automated web navigation, JavaScript execution, and data extraction using the cloud-based browser environment.
  • Sanitization: No mechanisms for filtering, sanitizing, or validating the external content are described in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 09:46 AM