cold-email-copy

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists exclusively of instructional Markdown files designed to guide an AI agent in copywriting and review tasks. No malicious patterns were identified across the 12 files.
  • [DATA_EXFILTRATION]: The skill instructions explicitly state that no API calls are made and that the agent should treat all provided facts as a closed set, precluding the exfiltration of sensitive data to external endpoints.
  • [COMMAND_EXECUTION]: No shell commands, system calls, or subprocess executions are present in the skill instructions or referenced resources.
  • [PROMPT_INJECTION]: The drafting and review instructions do not attempt to bypass safety guidelines or override system-level constraints. The review process is focused on quality and compliance with the user's brief.
  • [REMOTE_CODE_EXECUTION]: There are no patterns involving the download of external scripts, the installation of unverified packages, or the dynamic execution of code.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 12:34 AM
Security Audit — agent-trust-hub — cold-email-copy