rate-change-significance

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill includes a local Python script scripts/analyze.py for performing statistical tests (Z-test, Fisher's exact, McNemar). This script uses only standard libraries and does not perform any network operations, file system modifications outside its artifacts, or execution of untrusted code.\n- [SAFE]: The SQL generation process documented in references/sql-patterns.md includes explicit instructions for quoting identifiers and escaping values. It also mandates stopping and asking the user if a structured condition cannot be safely represented, which mitigates SQL injection risks.\n- [SAFE]: The skill uses official platform CLI tools (ae-cli) for metadata discovery and data extraction, following standard operational procedures for the target analytics environment.\n- [SAFE]: Analysis of the skill instructions and supporting files reveals no evidence of prompt injection, data exfiltration, or malicious persistence mechanisms. The metadata and instructions are consistent with the stated purpose of analyzing binary rate changes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 03:22 AM
Security Audit — agent-trust-hub — rate-change-significance