Drop into a skill folder
Personal: ~/.claude/skills/ab-test-learning-loop/SKILL.md
Project: .claude/skills/ab-test-learning-loop/SKILL.md
Brali Agent Skill · practical · usable
When a decision is reversible, compare two approaches around one clear question, keep the outcome you care about visible, and avoid treating a tiny informal comparison as statistical proof.
Portable instruction
Choose one reversible decision, define what you want to learn, compare two approaches as fairly as you can, and record the outcome before deciding what to try next. Use proper experimental design when the result needs statistical confidence.
What changed between the two approaches, what outcome did I observe, and is the evidence strong enough to act or only strong enough to suggest another test?
Install
Inspect SKILL.md first. The evidence state and skill mode are part of the contract.
Install for normal use.
Personal: ~/.claude/skills/ab-test-learning-loop/SKILL.md
Project: .claude/skills/ab-test-learning-loop/SKILL.md
hermes skills install https://brali-lifeos.github.io/skill-packs/ab-test-learning-loop/SKILL.mdWorkspace: <workspace>/skills/ab-test-learning-loop/SKILL.md
Global: ~/.openclaw/skills/ab-test-learning-loop/SKILL.md
Brali verification
Canonical ID: brali:ab-test-learning-loop
Evidence state: practical
Skill mode: usable
Recommendation eligible: yes
SHA-256: 75b97b65231da8b054264036358aac95a05be1e0ec3676016e04a1d6ecb0a34b
Brali CI verified. This means identity, state, mode, provenance and file integrity matched the generated corpus. It is not third-party vendor endorsement.
Brali keeps a one-to-one machine surface: one hack record, one stable skill identity. Trust state determines what the skill is allowed to do.
Original Brali knowledge is available for non-commercial reuse under CC BY-NC-SA 4.0. Keep attribution, canonical record link, evidence state and skill mode.