agent-browser
Pass
Audited by Gen Agent Trust Hub on May 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to install the tool (
npm i -g agent-browser) and initialize it (agent-browser install). It also relies on subsequent CLI commands for its core functionality. - [EXTERNAL_DOWNLOADS]: The skill downloads the
agent-browserpackage from the NPM registry. Furthermore, theagent-browser skills getcommand is used to dynamically fetch workflow instructions and command references from the vendor's remote repository at runtime. - [INDIRECT_PROMPT_INJECTION]: As a browser automation and scraping tool, this skill creates an attack surface for Indirect Prompt Injection. The agent is intended to process untrusted data from external websites, Slack conversations, and Electron apps, which could contain malicious instructions designed to manipulate the agent's behavior.
Audit Metadata