claude-in-chrome-expert

Pass

Audited by Gen Agent Trust Hub on Apr 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external web pages and provides guidance for interacting with it using powerful tools, creating a surface for indirect prompt injection.
  • Ingestion points: Tools such as read_page, get_page_text, read_console_messages, and read_network_requests (documented in SKILL.md and tools-reference.md) ingest arbitrary content from websites into the agent's context.
  • Boundary markers: The instructions do not provide or mandate the use of delimiters or 'ignore' instructions to wrap untrusted content retrieved from the browser.
  • Capability inventory: Documentation covers high-impact tools including javascript_tool for arbitrary JavaScript execution and form_input for element interaction, which could be manipulated by instructions hidden in page content.
  • Sanitization: There is no mention of sanitizing or validating content read from external pages before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 20, 2026, 11:27 AM
Security Audit — agent-trust-hub — claude-in-chrome-expert