INFORMATIVEACTIVEDocumentation Governance

Truth Source: Repository schemas and tests are authoritative.

Python SDK Guide

1. Purpose

The MPLP Python SDK (mplp-sdk) is the Reference implementation for Python. It provides a robust, schema-conformant implementation with full runtime support.

PyPI Package: mplp-sdk

2. Installation

pip install mplp-sdk

3. Features

• Protocol Conformance: Fully generated Pydantic v2 models from canonical JSON Schemas.
• Runtime Engine: Flexible ExecutionEngine supporting SA and MAP execution modes.
• Observability: Full Alignment with 12 Event Families (Truth Source). Built-in Validator and NDJSON Exporter.
  • Evidence: See project-governance/PDG_ALIGNMENT_EVIDENCE_REPORT_PY.md (Internal).
• Golden Flows: 5 verified protocol flows (Single Agent, Multi-Agent, Risk Confirm, Error Recovery, Network Transport).

4. Quick Start

4.1 Single Agent Execution

import asyncio
from mplp.model import Context, Plan
from mplp.validation import validate_context, validate_plan


context_data = {
    "title": "My Project",
    "domain": "code-assistance"
}
context = Context(**context_data)


is_valid = validate_context(context_data)


print(f"Context ID: {context.context_id}")

5. Golden Flows

Flow	Description	Test File
01	Single Agent	tests/test_flows_01_single_agent.py
02	Multi-Agent	tests/test_flows_02_multi_agent.py
03	Risk Confirmation	tests/test_flows_03_risk_confirm.py
04	Error Recovery	tests/test_flows_04_error_recovery.py
05	Network Transport	tests/test_flows_05_network_transport.py

6. Further Documentation

• packages/pypi/mplp-sdk/README.md
• packages/sources/sdk-py/docs/RUNTIME.md
• packages/sources/sdk-py/docs/OBSERVABILITY.md

7. Requirements

• Python >= 3.10
• pydantic >= 2.0, < 3.0