Adds statistical-analyst skill — fills a gap in the repo (no hypothesis
testing or experiment analysis tooling exists; only ab-test-setup for
instrumentation, but zero analysis capability).
Three stdlib-only Python scripts:
- hypothesis_tester.py: Z-test (proportions), Welch's t-test (means),
Chi-square (categorical) with p-value, CI, Cohen's d/h, Cramér's V
- sample_size_calculator.py: required n per variant for proportion and
mean tests, with power/MDE tradeoff table and duration estimates
- confidence_interval.py: Wilson score interval (proportions) and
z-based interval (means) with margin of error and precision notes
Validator: 86.4/100 (GOOD). Security audit: PASS (0 critical/high).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
c3693f9be1