skills/mukul975/anthropic-cybersecurity-skills/performing-brand-monitoring-for-impersonation/Gen Agent Trust Hub
performing-brand-monitoring-for-impersonation
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The brand monitoring logic uses
subprocess.runto execute thednstwisttool for domain permutation and typosquatting analysis. - Evidence:
cmd = ["dnstwist", "--registered", "--format", "json", ...]inSKILL.md. - The command is executed as a structured list of arguments without using a shell, mitigating command injection risks.
- [EXTERNAL_DOWNLOADS]: The skill interacts with external services to gather threat intelligence.
- Evidence: Requests to
safebrowsing.googleapis.com,api.twitter.com, andplay.google.cominSKILL.md. - Evidence: API calls to the specified target in
scripts/agent.py. - All external calls target well-known and trusted service providers or the user-defined target URL.
Audit Metadata