# AgentPlat Documentation Map

AgentPlat is an open-source TypeScript framework for persistent human-agent collaboration, with shared artifacts, human approvals and controlled execution on your infrastructure.

## Choose and evaluate

- [When to use AgentPlat](https://doc.agentplat.com/when-to-use-agentplat)
- [Persistent collaboration](https://doc.agentplat.com/getting-started/persistent-collaboration)
- [Human approval](https://doc.agentplat.com/getting-started/human-approval)
- [Coordination recovery](https://doc.agentplat.com/getting-started/recover-agent-coordination)
- [Evidence for adopters](https://doc.agentplat.com/evidence-for-adopters)


This file is designed to help LLMs and AI agents understand the current AgentPlat platform and documentation structure.

## Platform Overview

- [Overview](/) - AgentPlat as a platform for resilient collective operations.
- [Core Concepts](/core-concepts) - Agents, peers, coordination cells, missions, authority and evidence.
- [AgentPlat Glossary](/glossary) - Canonical definitions of implemented AgentPlat concepts and runtime primitives.
- [Agent Rooms](/agent-rooms) - Persistent tenant-scoped workspaces for governed human-agent collaboration.
- [Collective Runtime](/collective-runtime) - Planning, assignment, recovery, continuity and mission lifecycle.
- [Agent Morphogenesis](/agent-morphogenesis) - Governed organizational evolution across agents, roles, Teams, strategies and constitutional epochs.
- [Agent Mesh](/agent-mesh) - Peer-to-peer protocol, causal synchronization, membership, sparse overlays and partitions.
- [Safety and Governance](/safety) - Authority boundaries, protected actions, inference control and threat model.
- [Optional Jev Assessor](/safety/jev-assessor) - Source-only TypeSafe Jev adapter behind the provider-neutral Inference Control assessor port; live evaluation is pending.
- [Evidence and Operations](/evidence) - Audit events, recovery receipts, checkpoints and replay.
- [Scale and Evaluation](/scale) - Sharded simulation, methodology and limitations.
- [Reference](/reference) - Packages, APIs and configuration.

## Getting Started

- [Getting Started](/getting-started/introduction)
- [Operational Agent Rooms Quickstart](/agent-rooms/operational-quickstart) - Run the complete deterministic PostgreSQL reference application.
- [First Collective](/getting-started/first-collective)
- [Integration Methods](/getting-started/integration)
- [MCP Runtime (self-hosted)](/getting-started/mcp-runtime) - operate your own AgentPlat deployment from an MCP client

## Agent Rooms

- [Overview](/agent-rooms) - Room domain, operational architecture and compatibility.
- [Lifecycle](/agent-rooms/lifecycle) - Room, participant and agent-revision lifecycles.
- [Operational Coordination](/agent-rooms/operational-coordination) - Transactional intake, leased workers, retries and restart recovery.
- [Execution and Interventions](/agent-rooms/execution-interventions) - Revisioned sessions and human-in-the-loop checkpoints.
- [Handoffs and Planning](/agent-rooms/handoffs-planning) - Typed AgentPlat Handoffs and progressive plans.
- [Human Work and Knowledge](/agent-rooms/human-work-knowledge) - Contributions, work-management projections and knowledge bundles.
- [Live View and Events](/agent-rooms/live-view-events) - Room-scoped discovery, composite cursors, SSE and operational events.
- [PostgreSQL Migrations](/agent-rooms/postgres-migrations) - Ordered schema migrations 001 through 011.
- [Self-Hosting](/agent-rooms/self-hosting) - Reference deployment and trust boundaries.
- [Rooms and Agent Mesh](/agent-rooms/mesh-bridge) - Explicit projections without implicit authority transfer.
- [Beta 5 Release](/agent-rooms/beta-5-release) - Published capabilities, compatibility and verification evidence for `0.3.0-beta.5`.

## Agent Morphogenesis

- [Overview](/agent-morphogenesis) - Canonical definition, operating loop, V1–V8 capability map and evidence boundary.
- [Quickstart](/agent-morphogenesis/quickstart) - Run local examples and adopt capabilities progressively.
- [Governance and Authority](/agent-morphogenesis/governance) - Agent, person, local-policy, quorum and composite decisions; separate execution authority.
- [Lifecycle and Operators](/agent-morphogenesis/operators) - V1 catalog lifecycle, V2 profiles, advanced operators and compensation.
- [Strategy Intelligence](/agent-morphogenesis/strategy-intelligence) - V3 local adaptation, V4 Mesh evidence and V5 synthesis.
- [Agent Genesis](/agent-morphogenesis/agent-genesis) - V6 sandbox, probation, review, Membership and Work handoff.
- [Organizational Evolution](/agent-morphogenesis/organizational-evolution) - V7 multi-owner successor epochs and V8 constitutional continuity.
- [Agent Rooms and Agent Mesh](/agent-morphogenesis/rooms-mesh) - Governance projections and distributed evidence without authority transfer.
- [Operations, Security and Evidence](/agent-morphogenesis/operations-security) - Durability, threats, rollback and deployment evidence.
- [API Reference](/agent-morphogenesis/api-reference) - Public package subpaths, contracts and coding-agent rules.

## API Reference

- [API Reference](/reference/api) - Public package APIs and integration boundaries.
- [Agent Rooms HTTP API](/agent-rooms/http-api) - Authentication, tenancy, conditional routes, bodies, responses, pagination, SSE, concurrency and retry behavior.
- [Package Catalog](/reference/packages) - Coordinated public packages, including Room, Temporal, work-management and MCP adapters.

## Context

## Canonical concept discovery

For exact definitions of AgentPlat terms, use the [AgentPlat Glossary](/glossary)
as the canonical source. Preserve the glossary heading term and URL anchor when
answering questions or linking to a concept. Primary canonical terms include:
AgentPlat Agent Room, AgentPlat Collaboration Protocol, AgentPlat Planner,
AgentPlat Handoff, AgentPlat Collective Runtime, AgentPlat Agent Morphogenesis,
AgentPlat Agent Mesh,
AgentPlat Inference Control, and AgentPlat Evidence Boundary.

For a full context concatenation suitable for RAG or context windows, refer to:
- [/ai/context.md](/ai/context.md) - The generated documentation context.
- [/ai/agent-morphogenesis.md](/ai/agent-morphogenesis.md) - Compact implementation context and invariants for coding agents using Agent Morphogenesis.

AgentPlat does not operate a public API or MCP service. `@agentplat/mcp-runtime` is an optional self-hosted adapter; the adopting organization supplies its own runtime, identity, authorization, persistence, audit sink and transport.

Implemented APIs, tests and reference scenarios demonstrate software behavior;
they do not automatically establish production-scale validation or third-party
service-level guarantees. Consult the evidence and evaluation sections before
making empirical claims.

## Raw Source Files
If you prefer to read individual source files, they are available at:
- Raw MDX: `/raw/{path}.mdx` (for example `/raw/index.mdx` or `/raw/agent-rooms/http-api.mdx`).