skills/leeguooooo/chrome-use/electron/Gen Agent Trust Hub

electron

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: Instructions to launch desktop applications with the --remote-debugging-port flag, which exposes the application internal state via the Chrome DevTools Protocol and grants the agent programmatic control.
  • [DATA_EXFILTRATION]: Provides patterns to extract UI data and application state using chrome-use get text and chrome-use snapshot --json, specifically targeting sensitive applications such as Slack and 1Password.
  • [PROMPT_INJECTION]: Vulnerability surface for indirect prompt injection. 1. Ingestion points: Data entering context via chrome-use snapshot and chrome-use get text (SKILL.md). 2. Boundary markers: Absence of delimiters to distinguish application content from agent instructions. 3. Capability inventory: Use of Bash(chrome-use:*) for interaction (SKILL.md). 4. Sanitization: No sanitization of extracted content before processing.
  • [EXTERNAL_DOWNLOADS]: Neutral documentation of npx chrome-use which fetches and executes the automation tool from the npm registry.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 03:03 AM
Security Audit — agent-trust-hub — electron