browserbase-cli
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for the global installation of the
browseCLI package usingnpm install -g browse. - [REMOTE_CODE_EXECUTION]: The command
browse skills installinitiates remote code execution by usingnpxto download and run installer scripts. - [COMMAND_EXECUTION]: The skill uses the
Bashtool to perform platform operations, such as creating sessions and managing functions through thebrowseCLI. - [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface by incorporating untrusted data from the web into the agent's context. \n
- Ingestion points: Web content is ingested through the
browse cloud fetch,browse cloud search, andbrowse getcommands as described inSKILL.mdandREFERENCE.md. \n - Boundary markers: There are no boundary markers or instructions to treat external content as untrusted data within the provided scripts or instructions. \n
- Capability inventory: The skill allows for command execution via
Bash, filesystem writes using--outputflags, and network operations. \n - Sanitization: No sanitization or content filtering is specified for the data retrieved from external sources.
Audit Metadata