search-intent-coverage

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides explicit shell commands for the agent to execute, including environment setup (npm i -g agent-browser) and operational commands (agent-browser open, snapshot, etc.) to interact with web browsers.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the installation of the agent-browser package from the public npm registry to enable its web-scraping capabilities.
  • [DATA_EXFILTRATION]: The core workflow involves fetching and extracting text/data from external competitor URLs. This is the intended purpose of the skill, though it involves automated navigation to third-party domains.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from external websites.
  • Ingestion points: Untrusted content is ingested from competitor URLs via the agent-browser tool.
  • Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions embedded within the scraped competitor content.
  • Capability inventory: The skill environment has shell execution capabilities and browser control.
  • Sanitization: No evidence of sanitization or filtering is present for the content extracted from the web before it is analyzed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 06:28 PM