Skills
Navigate
skills/am-will/codex-skills/gemini-computer-use/Gen Agent Trust Hub

gemini-computer-use

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted web content which could contain malicious instructions for the agent model.
  • Ingestion points: The agent in scripts/computer_use_agent.py navigates to external URLs via page.goto() and interacts with the page content.
  • Boundary markers: No delimiters or safety instructions are provided to separate user prompts from the content of the visited websites.
  • Capability inventory: The script scripts/computer_use_agent.py provides the agent with extensive control capabilities including type_text_at, click_at, and key_combination.
  • Sanitization: There is no evidence of content sanitization or filtering before the agent processes web page screenshots and URLs.
  • [COMMAND_EXECUTION]: The skill executes browser-level automation commands such as navigation, mouse clicks, and keyboard inputs based on LLM-generated function calls.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 11:10 PM