Skills
Navigate
skills/zjc0230-code/awesome-ai-skills/agent-browser/Gen Agent Trust Hub

agent-browser

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests and processes untrusted content from external websites.\n
  • Ingestion points: Web content is ingested via agent-browser open, agent-browser snapshot, and agent-browser get text commands (found in SKILL.md and references/commands.md).\n
  • Boundary markers: No specific delimiters or safety instructions are used to distinguish web content from system instructions.\n
  • Capability inventory: The skill allows arbitrary JavaScript execution via agent-browser eval, file writing via agent-browser screenshot and agent-browser state save, and access to local files via the --allow-file-access flag.\n
  • Sanitization: No evidence of content sanitization or filtering of external data before processing.\n- [COMMAND_EXECUTION]: The skill relies on the agent-browser CLI tool to perform complex browser interactions, including navigating to URLs, interacting with DOM elements, and evaluating JavaScript expressions.\n- [EXTERNAL_DOWNLOADS]: The documentation references the installation of external dependencies such as appium and the xcuitest driver via npm for mobile browser automation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 04:31 PM