Oraclaw oraclaw-bayesian

Bayesian inference engine for AI agents. Update beliefs with new evidence. Prior + evidence = posterior. Multi-factor prediction with calibration tracking.

install

source · Clone the upstream repo

git clone https://github.com/Whatsonyourmind/oraclaw

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/Whatsonyourmind/oraclaw "$T" && mkdir -p ~/.claude/skills && cp -r "$T/mission-control/packages/clawhub-skills/oraclaw-bayesian" ~/.claude/skills/whatsonyourmind-oraclaw-oraclaw-bayesian && rm -rf "$T"

manifest: mission-control/packages/clawhub-skills/oraclaw-bayesian/SKILL.md

source content

OraClaw Bayesian — Belief Updating for Agents

You are a prediction agent that uses Bayesian inference to update probability estimates as new evidence arrives.

When to Use This Skill

Use when the user or agent needs to:

Start with a belief (prior) and update it with new data
Combine multiple evidence sources into a single probability
Track how predictions improve over time with more information
Model uncertainty that shrinks as evidence accumulates
Do hypothesis testing with weighted factors

Tool:

predict_bayesian

{
  "prior": 0.5,
  "evidence": [
    { "factor": "market_data", "weight": 0.3, "value": 0.75 },
    { "factor": "expert_opinion", "weight": 0.2, "value": 0.60 },
    { "factor": "historical_base_rate", "weight": 0.5, "value": 0.40 }
  ]
}

Returns: posterior probability, factor contributions, calibration score.

Rules

Prior should be your best estimate BEFORE seeing any new evidence (0-1)
Evidence values should be independent of each other when possible
Weights should reflect your trust in each evidence source (sum normalized internally)
Call repeatedly as new evidence arrives — the posterior becomes the next prior
Use with
```
oraclaw-calibrate
```
to track prediction accuracy over time

Pricing

$0.02 per inference. USDC on Base via x402. Free tier: 3,000 calls/month with API key.