authoring-log-alerts
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows secure patterns by using a local, standard-library-only Python script (
scripts/baseline_stats.py) to perform complex mathematical calculations on tool outputs. This prevents LLM hallucinations and ensures data stays within the execution environment. - [COMMAND_EXECUTION]: The instructions involve executing a local script and a Python one-liner for threshold calculation. Both use standard libraries, process JSON data safely, and do not involve remote code execution or shell injection risks.
- [DATA_EXPOSURE]: There are no hardcoded credentials or attempts to access sensitive system files. The skill interacts exclusively with PostHog's logging and alerting tools. It includes explicit instructions to confirm notification destinations with the user before attaching alerts to channels, preventing accidental data disclosure to public or incorrect channels.
Audit Metadata