paw-mkt-guerrilla
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install the
agent-browsertool from Vercel Labs' official GitHub repository. Vercel Labs is a well-known and trusted technology provider. - [COMMAND_EXECUTION]: The skill utilizes shell commands for environment setup and tool discovery, including running local scripts like
tool-discovery.shandchrome-profiles.sh. It also suggests the use ofnpmandnpxfor package management and tool execution. - [DATA_EXFILTRATION]: The skill employs
agent-browserfor performing research on live websites, which involves network operations. The documentation includes best practices for security, such as adding authentication state files to.gitignoreand using dedicated browser profiles to prevent accidental credential leakage. - [PROMPT_INJECTION]: The skill performs 'Live Website URL Audits' using
agent-browserorWebFetch. This creates a surface for indirect prompt injection where untrusted content from external websites is ingested into the agent's context. The skill mitigates this by grounding actions in pre-flight strategic context files and using structured workflows.
Audit Metadata