Explore in MochiBot. Install in chat.Hosted workspaces and MochiBot lightweight setups include the Skills assistant. Choose a Skill and ask for installation in chat. Other OpenClaw setups can follow the first-install guide.
How to use Skills →
MOCHILAB / AI SKILL
A/B Conversion Evidence
Compare two independent binary conversion samples with Fisher's exact test, Wilson intervals and explicit validity limits.
What it helps with
Compare two independent binary conversion samples with Fisher's exact test, Wilson intervals and explicit validity limits.
Say it like this
Compare 6 of 300 conversions with 12 of 300 and explain the uncertainty.
We peeked during the test; separate descriptive results from valid inference.
What you receive
Ask whether each unit was independently randomized once, whether the binary outcome window is complete, and whether the fixed endpoint was preplanned without optional stopping. Require exactly two arms; report observed rates, absolute change, Fisher two-sided p and Wilson 95% intervals. State that p is not a winner probability and that significance is not practical value.
Give the assistant your actual files or records and confirmed rules. It asks for missing facts rather than inventing records; you do not need to hand-fill JSON. Receive a traceable report and detailed review rows for download.
Scope
Two independent binary arms, one declared outcome. No revenue, clustered observations, multiple comparisons, sequential design, causal claim from observational groups or automatic rollout.
The script reads supplied local files and creates reports without network or model calls. Your assistant still uses its configured services; this does not promise absolute privacy for a conversation. The Skill is free and needs Python3.10+. Optional AI interpretation follows your existing model service charges.
Demand and technical basis