dev-browser
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install the
dev-browserpackage from a public registry and notes that the utility may download a browser binary during its first run. - [COMMAND_EXECUTION]: The skill involves executing shell commands to install the
dev-browsertool and to interact with it, including passing JavaScript code via stdin and CLI flags. - [INDIRECT_PROMPT_INJECTION]: The skill provides the agent with capabilities to browse and scrape external web pages, which serves as a surface for indirect prompt injection.
- Ingestion points: Data from untrusted external sources enters the agent's context through methods such as
page.gotoandpage.snapshotdefined in theSKILL.mdinstructions. - Boundary markers: The instructions do not define specific delimiters or warnings to separate external web content from the agent's internal instructions.
- Capability inventory: The skill enables full browser control including clicking, form filling, and data extraction, as well as executing generated JavaScript via the CLI.
- Sanitization: There are no documented procedures for sanitizing or filtering data retrieved from the web before it is interpreted by the agent.
- [DYNAMIC_EXECUTION]: The skill relies on the agent dynamically generating and executing JavaScript snippets at runtime to control the browser automation process.
Audit Metadata