Governance Overview
MPLP is an open standard. Its evolution is governed by a normative RFC process, transparent versioning rules, and a commitment to stability.
Bounds
Scope Boundary
MPLP defines lifecycle semantics for interoperability while maintaining strict neutrality towards implementation details.
✓ Agent Lifecycle
Initialization, plan creation, and trace recording.
✓ Governance Semantics
Permission boundaries and coordination rules.
✓ Observability
Audit trails and lifecycle state transitions.
✓ Interoperability
Vendor-neutral coordination schemas.
✗ Model Training
How LLMs are trained or fine-tuned.
✗ Certification
MPLP does not issue certification badges.
✗ Prompt Engineering
Internal prompt structures or optimization.
✗ Runtime Execution
Governs semantics; runtimes execute code.
Process
How We Evolve
Changes to the protocol follow a strict Request for Comments (RFC) process. Anyone can propose a change, but it must pass through stages of review before adoption.
Draft
Initial proposal (RFC). Open for community discussion, feedback, and red-lining.
Review
Formal review by core maintainers. Normative changes are mutually agreed upon and refined.
Frozen
Spec is locked and versioned. No breaking changes allowed without a major version bump.
Versioning Rules
Protocol Status & Context
Current Status
v1.0.0 — Protocol Frozen
Freeze date: 2025-12-03
Exploration Areas (Non-Binding)
The following areas are subjects of potential future exploration, subject to community interest and governance approval. No commitments are made.
- →Multi-Agent CollaborationRefining negotiation patterns and handoff protocols.
- →Learning & OptimizationDefining feedback loops and interaction surfaces.
- →Enterprise ExtensionsFormalizing authority chains and auditability standards.
Positioning & Drift Control
Governance Evidence
Website governance artifacts, baselines, and CI evidence are maintained in the protocol repository.
This website provides discovery and positioning content only. Normative requirements live in Docs; the ultimate source of truth is the Repository.
Evidence is evaluated through documentation-defined scenarios and replayable artifacts. This website does not certify or endorse implementations.