Model Context Protocol (MCP) Server

ALP natively supports Anthropic's Model Context Protocol, allowing modern AI IDEs (like Claude Desktop, Cursor, and Windsurf) to securely query your .alp workspace and understand your architecture in real time.

Installation

The MCP server is provided via the @autonomous-lifecycle-protocol-alp/mcp-server package.

npm install -g @autonomous-lifecycle-protocol-alp/mcp-server

Available Tools

Once connected, your IDE gains access to the following tools:

Core Tools

Tool	Description
alp_list_objects	List all objects, optionally filtered by type (e.g., task, agent, memory)
alp_read_object	Read a specific ALP object by its ID to get detailed instructions, rules, or memory
alp_get_graph	Get the full dependency graph of the ALP workspace as a sorted execution order
alp_get_status	Get the current project status (task counts by state: [x], [~], [ ], [!], [?])
alp_validate	Validate the ALP workspace to ensure the agent hasn't introduced syntax errors
alp_update_status	Update the status of a task (supports the [?] review marker for HITL handoffs)
alp_get_impact	Get all downstream nodes affected by a change to a given node
alp_search	Fuzzy search across all object IDs and descriptions
alp_delegate	Create a new task assigned to a specific role/agent
alp_decompose	Split a large task into sub-tasks blocked by the parent
alp_create_task	Scaffold a new task .alp file
alp_create_feature	Scaffold a new feature .alp file

Events & Analytics

Tool	Description
alp_get_events	Read recent runtime events with optional type filtering and limit
alp_get_analytics	Analytics summary from state store or event log

Governance & Registry — v4+

Tool	Description
alp_check_policy	Evaluate policy decisions — check if a file path or command is permitted by workspace @policy rules
alp_visualize	Generate a Mermaid DAG diagram of the project dependency graph for embedding in docs or chat
alp_search_registry	Search installed ALP registry packages
alp_get_timelines	Retrieve @timeline scheduling objects and their cron/at triggers

Contracts & Vaults

Tool	Description
alp_get_contracts	List all @contract objects and their allow/deny rules
alp_get_vaults	List all @vault objects and their recipient/algorithm metadata

Swarm Marketplace & Event Mesh — v38.0.0

Tool	Description
alp_get_swarm_marketplace	List registered skills from @swarm_marketplace objects, optionally filtered by category
alp_get_event_mesh	List event mesh topics and recent events from @event_mesh objects

Macro Engine — v37+

Tool	Description
alp_get_macros	List @macro definitions from the workspace
alp_expand_macro	Expand a @macro definition by ID and return generated objects

Memory Mesh — v38+

Tool	Description
alp_memory_store	Store a memory node in the workspace memory mesh
alp_memory_query	Query the memory mesh for relevant memories
alp_memory_stats	Return memory mesh statistics

In addition, the server exposes the workspace .alp files as MCP resources (file:// URIs) so a client can read raw object files directly.

Usage with Claude Desktop

To add ALP to Claude Desktop, edit your claude_desktop_config.json and add the server:

{
  "mcpServers": {
    "alp": {
      "command": "alp-mcp",
      "args": []
    }
  }
}

Usage with Cursor

Add to your .cursor/mcp.json in your project root:

{
  "mcpServers": {
    "alp": {
      "command": "alp-mcp",
      "args": [],
      "env": {}
    }
  }
}

Usage with Windsurf

Add to your Windsurf MCP configuration (~/.windsurf/mcp.json):

{
  "mcpServers": {
    "alp": {
      "command": "alp-mcp",
      "args": []
    }
  }
}

Now, your AI coding assistant can natively read your project's architecture, enforce policies, visualize dependencies, and query task states before writing a single line of code!