Project

You manage project knowledge with a structured MCP server that models projects, tasks, milestones, resources, and team interactions as a knowledge graph. This server keeps context across planning sessions, tracks progress, and helps you reason about risk, dependencies, and resource allocation so you can make informed decisions.

How to use

You interact with the MCP server through a client that can start sessions, load detailed context for entities, and record session results. Start a new project management session to review your project landscape. Load context for a specific project, task, milestone, or other entity to see its current status, priority, and any sequential relationships. During or after a session, end the session with a structured summary that captures achievements, task updates, new tasks, status changes, and overall project health. Use the domain functions to get an overview of a project, analyze task dependencies, assess risks, and review the decision history.

Key practical patterns you can follow:

• Start a session to get a live view of active statuses, priorities, and upcoming work.
• Load context for a project or task to drill into its details, including where it sits in the sequence (precedence) and its current status.
• Build context to add new tasks, assign owners, set priorities, and establish dependencies.
• End a session with a clear summary, noting achievements and changes to the project health or milestones.
• Use advanced queries to inspect the full knowledge graph, find related entities, or identify high-priority work and critical paths.

How to install

{
  "mcpServers": {
    "project": {
      "type": "stdio",
      "name": "project_mcp",
      "command": "npx",
      "args": ["-y", "github:tejpalvirk/project"]
    }
  }
}

You can also run the MCP server locally by installing the package globally and starting with a direct command. The following example shows the runtime command you use after installation. This config uses a local executable name to start the server.

{
  "mcpServers": {
    "project": {
      "type": "stdio",
      "name": "project_mcp",
      "command": "contextmanager-project",
      "args": []
    }
  }
}

If you prefer to run the server via Docker, use the Docker configuration snippet to start a container that runs the MCP server image. This keeps the server isolated and easy to manage.

{
  "mcpServers": {
    "project": {
      "type": "stdio",
      "name": "project_mcp",
      "command": "docker",
      "args": ["run","--rm","-i","mcp/project"]
    }
  }
}

Configuration and environment

The server supports environment variables to control data storage and session history. Set the paths for the knowledge graph and sessions before you start the service.

MEMORY_FILE_PATH="./pm-memory.json" SESSIONS_FILE_PATH="./pm-sessions.json" npx github:tejpalvirk/contextmanager-project

• MEMORY_FILE_PATH: Path where the knowledge graph data will be stored. Can be absolute or relative. Default is ./project/memory.json.
• SESSIONS_FILE_PATH: Path where session data will be stored. Can be absolute or relative. Default is ./project/sessions.json.

Examples and prompts

Start a new session and review current state:

• Let’s start a new project management session to review the Mobile App Development project.

Load and inspect a specific context:

• Load the context for the Mobile App Development project so I can see its current status.

Record session results and updates:

• I’ve just finished a project review meeting for Mobile App Development. We completed the UI design milestone, identified 2 new risks, and assigned 3 new tasks. The UI tasks are complete, API tasks are high priority, and we should monitor API risks closely.

Security and notes

Treat session data and the knowledge graph as confidential project information. Use role-appropriate access and rotate credentials as needed. Log important decisions and keep an audit trail to understand how project health and priorities evolve over time.

Troubleshooting

If you encounter startup issues, verify your storage paths exist and are writable. Check that the command and arguments used to start the server exactly match the runtime environment you’re using (npx, npm global install, or Docker). Ensure the MCP client you use can connect to the local stdio server and that the environment variables for data paths are set correctly.

Notes

The server exposes a comprehensive set of domain functions to retrieve and analyze project data, including health, risks, milestones, timeline, and dependencies. Use these functions to build reports, evaluate alternatives, and drive decision-making across your projects.

Available tools

getProjectOverview

Provides a comprehensive view of a project including tasks, milestones, team members, issues, risks, and overall health.

getTaskDependencies

Analyzes task dependencies to identify blocked tasks, critical paths, and opportunities to re-sequence work.

getTeamMemberAssignments

Retrieves all assignments for a specific team member, showing workload and task status.

getMilestoneProgress

Tracks progress toward key project milestones and highlights any slippage.

getProjectTimeline

Analyzes project timelines with key dates and critical path insights.

getResourceAllocation

Examines how resources are allocated across projects and tasks.

getProjectRisks

Identifies and assesses project risks with context and mitigations.

findRelatedProjects

Discovers connections between different projects and shared resources or dependencies.

getDecisionLog

Tracks decision history and context to support traceability.

getProjectHealth

Assesses overall project health with metrics and recommendations.

getStatusOverview

Displays all entities with a specific status value (inactive, active, complete).

getPriorityItems

Identifies high-priority tasks and activities needing attention.

getTaskSequence

Visualizes the sequence of tasks based on precedes relations.