v3.1 Release

What's New in
Product Org OS

Discovery and learning skills: validate assumptions before committing, and compound learnings after shipping. The Vision to Value loop gets sharper.

4 new skills. 65 total. Stronger Phase 1 and Phase 6.

65
Skills
+4
New in v3.1
3
Discovery
1
Learning
🔍
Discovery

Validate Before
You Commit

Three new Phase 1 skills bring structured discovery into the Vision to Value flow. Map assumptions, explore opportunity spaces, and design experiments before resources are committed. Rooted in the work of Teresa Torres, Itamar Gilad, and Alberto Savoia.

  • ✓ Map and prioritize assumptions with impact vs. uncertainty scoring
  • ✓ Build opportunity solution trees from desired outcomes down to testable solutions
  • ✓ Design experiments: Fake Door, Concierge, Wizard of Oz, and 5 more types
  • ✓ Auto-routing: mention "assumption", "experiment", or "validate" and the PM handles it
New in v3.1

The New Skills

Each skill generates structured outputs and integrates with the context layer.

Phase 1 · Discovery
/assumption-map
Identify hidden assumptions behind a strategy or bet. Score each by impact and uncertainty, plot on a 2×2 matrix, and plan validation for the riskiest ones first.
Based on Itamar Gilad's Assumption Mapping
Phase 1 · Discovery
/opportunity-tree
Start from a desired outcome, discover opportunity spaces through customer research, structure them into a navigable tree, and generate solution ideas with experiment plans.
Based on Teresa Torres' Continuous Discovery Habits
Phase 1-2 · Validation
/experiment-design
Design lean experiments with clear hypotheses and success criteria. Choose from 8 experiment types: Fake Door, Concierge, Wizard of Oz, Smoke Test, and more.
Based on Savoia's Pretotyping methodology
Phase 6 · Learning
/compound
Extract insights from outcomes, perform root cause analysis, generalize findings into reusable principles, and save to the context layer so the org gets smarter.
Closes the Vision to Value learning loop
🧪
Learning Loop

Learnings That
Compound

Phase 6 gets its missing piece. The /compound skill transforms raw outcome data into reusable organizational intelligence. Extract what happened, understand why, and generalize into principles that improve future decisions.

  • ✓ 5-step workflow: Extract, Root Cause, Generalize, Save, Index
  • ✓ Auto-saves to context/learnings/ for cross-session recall
  • ✓ Feeds back into Phase 1 through /relevant-learnings
New in v3.1
🔄
Vision to Value Loop

Stronger Where It
Matters Most

Product work fails at the edges: teams skip discovery and rush to build, or ship and never look back. v3.1 strengthens exactly those two phases. Phase 1 gets structured discovery. Phase 6 gets systematic learning extraction. The Vision to Value loop tightens.

  • ✓ Phase 1: 5 skills → 8 skills (discovery + validation added)
  • ✓ Phase 6: 5 skills → 6 skills (learning extraction added)
  • ✓ Routing updated: "assumption", "experiment", "validate" auto-route to @pm
  • ✓ All existing skills and agents unchanged. Fully backward compatible.
Enhanced in v3.1

Get Started with
v3.1

Install in one command. 13 agents, 65 skills, and the full Vision to Value operating system ready in seconds.

$ claude plugins install github:yohayetsion/product-org-os
GitHub ← Product Org OS Home
What was new in v3.0