Skip to content

Quick Start

Jump to bottom
Roberton003 edited this page Jul 22, 2026 · 2 revisions

Quick Start (5 Minutes)

Install

pip install mcp-server-decisions

Configure

Add to your MCP client config (e.g., Claude):

{
  "mcp-server-decisions": {
    "type": "stdio",
    "command": "mcp-server-decisions"
  }
}

Record Your First Decision

Ask your agent:

"Record this decision: we're using DuckDB for caching because it's fast and requires no external infra. Alternatives: Redis, SQLite, Elasticsearch. Predict p99 latency will drop below 200ms and storage cost will stay under $50/month."

The agent records:

  • Problem: Query latency exceeds SLA
  • Solution: DuckDB caching
  • Rejected: Redis, SQLite, Elasticsearch
  • Technologies: duckdb
  • Predictions: LATENCY (p99 < 200ms), COST (< $50/month)

Returns: DEC-2026-0001 with prediction IDs

Later: Record Outcome

Once you've run the system:

"We measured p99 latency at 180ms with DuckDB. Record this as 95% accurate."

Agent records:

  • Actual value: p99 = 180ms
  • Accuracy: 95 (SUCCESS ✅)

System automatically validates and stores the outcome.

Check Performance

python3 -m pip install -e .
python3 scripts/technology_performance_report.py

Output:

technology: duckdb | successful: 1 | failed: 0 | avg_accuracy: 95.0% | confidence: LOW

Key Concepts

Term Meaning
Decision ID DEC-YYYY-NNNN uniquely identifies a choice
Prediction ID PRD-YYYY-NNNN links a measurable claim to a decision
Accuracy Score 0-100: how close actual result was to predicted
Outcome Gate Nudge in responses reminding you to close loops

Next Steps

Tip: The Outcome Gate field appears automatically in tool responses when you have pending predictions. It keeps you from forgetting to validate decisions.

Clone this wiki locally