The Yao Bayesian Decision Skill transforms complex, uncertain decisions into structured, iterative judgment processes. Instead of asking AI for a single answer, it guides you through multi-round dialogue where each round updates probability estimates based on new evidence. The output is a bilingual (Chinese/English) Markdown + HTML decision report showing the reasoning trail, judgment changes, and recommended actions.
Source: GitHub - yaojingang/yao-open-skills
How It Works
Incomplete input (your messy, real-world question)
↓
Initial assessment (establish weak prior assumptions)
↓
Prior selection (pick 3-5 relevant heuristics from 20 judgment principles)
↓
┌────────────────────┐
│ Iterative rounds: │
│ Ask questions │ ← Multi-round dialogue
│ Gather evidence │
│ Update estimates │
│ Track what changed │
└────────┬───────────┘
↓
Decision readiness check (confidence level assessment)
↓
Final report (Markdown + bilingual HTML)
The key insight: it doesn’t wait for perfect information. It starts from incomplete input, builds weak priors, and iteratively strengthens the judgment through structured questioning — tracking exactly which evidence changed which estimates.
What Makes It Different
| Traditional AI Decision | Bayesian Skill |
|---|---|
| Single prompt → single answer | Multi-round dialogue → evolving judgment |
| Black box reasoning | Evidence attribution (which fact changed which estimate) |
| Assumes complete info | Starts from incomplete input |
| One format | Bilingual Markdown + interactive HTML |
| Static | Iterative refinement with confidence tracking |
Output Format
The skill generates:
- Markdown report — decision process, judgment changes, and action items
- Bilingual HTML (Chinese/English) — interactive with sticky navigation, language switching
- Print-friendly PDF — exportable from browser
- Evidence trail — which round introduced which information and how it shifted the judgment
Use Cases
The skill is designed for decisions where information is incomplete and risk is non-trivial:
- Product decisions — Should we ship this feature this quarter?
- Growth strategy — Which market should we enter first?
- Career choices — Should I take this job offer?
- Investment decisions — Is this startup worth investing in?
- Personal decisions — Should I relocate? Buy or rent?
- Research direction — Which hypothesis is worth pursuing?
Installation
# Add to Claude Code via skills protocol
npx skills add yaojingang/yao-open-skills
How LearnAI Team Could Use This
- Teaching decision analysis — the iterative Bayesian framework is itself a pedagogical tool
- Research planning — use it to systematically evaluate which research directions to pursue
- Grant decisions — structure the “should we apply for this grant?” analysis
- Student advising — help students work through career and academic decisions systematically
Real-World Use Cases
- Startup founders — structured pivot vs persist decisions
- Product managers — feature prioritization under uncertainty
- Researchers — hypothesis selection with explicit evidence tracking
- Anyone facing a complex decision — the process forces you to articulate what you know, what you don’t, and what would change your mind