foundations-causal-inference

Pass

Audited by Gen Agent Trust Hub on Oct 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides educational templates and methodological guidance for causal inference (e.g., Difference-in-Differences, Instrumental Variables, Synthetic Control). The content is academic and informational, with no instructions to bypass security filters or exfiltrate data.
  • [COMMAND_EXECUTION]: The file scripts/test_evalue.py contains calls to subprocess.run(). This is used exclusively for unit testing the CLI interface of the included evalue.py script. The execution is limited to the local Python interpreter and specific, hardcoded arguments for validation purposes, posing no risk of command injection or malicious execution.
  • [SAFE]: All external URLs referenced in the skill's templates and data/sources.json target reputable academic publishers (e.g., Oxford University Press, Econometrica), pre-print servers (ArXiv, NBER), and official technology organizations (PyWhy, Google). No suspicious or obfuscated links were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 17, 2026, 08:26 AM