--- name: tensorpm description: "AI-powered project management - a Notion and Jira alternative with local-first architecture. Manage projects, track action items, and coordinate teams via MCP tools or A2A agent communication. Signed & notarized. https://tensorpm.com" compatibility: Requires TensorPM desktop app to be running for MCP tools and A2A communication. Available on macOS, Windows, and Linux. metadata: homepage: https://tensorpm.com author: tensorpm team ---

TensorPM Skill

AI-Powered Project Management - Intelligently manage projects, track action items, and coordinate teams with context-driven prioritization.

Local-first, no account required. Full app, free forever — use your own API keys (OpenAI, Claude, Gemini, Mistral) or local models (Ollama, vLLM, LLM studio). Optional: Account with cloud sync enables E2E encrypted collaboration across devices and teams.

Interact with TensorPM via MCP tools or A2A agent communication.

Signed & Notarized: macOS builds are code-signed and notarized by Apple. Windows builds are signed via Azure Trusted Signing.

Download

macOS (Homebrew)

brew tap neo552/tensorpm
brew install --cask tensorpm

Linux (Terminal)

curl -fsSL https://tensorpm.com/download/linux -o ~/TensorPM.AppImage
chmod +x ~/TensorPM.AppImage

Direct Downloads

• Windows: TensorPM-Setup.exe
• macOS: TensorPM-macOS.dmg
• Linux: TensorPM-Linux.AppImage

Release Notes:

Alternative:

Setup

MCP Integration (Automatic)

TensorPM includes a built-in MCP server that runs locally. Install from within the app:

1. Open TensorPM
2. Go to Settings → Integrations
3. Click Install for your AI client

Requirement: TensorPM must be running for MCP tools to work.

Setting AI Provider Keys via MCP

Use the set_api_key tool to configure AI providers directly from your AI client:

set_api_key
  provider: "openai"      # openai, anthropic, google, mistral
  api_key: "sk-..."

Keys are securely stored in TensorPM. Write-only - keys cannot be read back.

A2A Configuration

TensorPM exposes a local A2A agent endpoint on port 37850.

No authentication required — A2A runs on localhost only, all local requests are trusted.

Agent Discovery

Step 1: Get Root Agent Card

curl http://localhost:37850/.well-known/agent.json

Returns the root agent card with links to all project agents.

Step 2: List Projects

curl http://localhost:37850/projects

Returns:

[
  {
    "id": "project-uuid",
    "name": "My Project",
    "agentUrl": "http://localhost:37850/projects/project-uuid/a2a",
    "agentCardUrl": "http://localhost:37850/projects/project-uuid/.well-known/agent.json"
  }
]

Step 3: Get Project Agent Card

curl http://localhost:37850/projects/{projectId}/.well-known/agent.json

Returns the A2A agent card for a specific project with capabilities and supported methods.

Talking to a Project Agent

Send messages to a project's AI agent using JSON-RPC:

curl -X POST http://localhost:37850/projects/{projectId}/a2a \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "message/send",
    "id": "1",
    "params": {
      "message": {
        "role": "user",
        "parts": [{"kind": "text", "text": "List high-priority items"}]
      }
    }
  }'

Supported JSON-RPC methods:

| Method | Description | |--------|-------------| | message/send | Send a message and get a blocking response | | message/stream | Send a message and stream the response via SSE | | tasks/get | Retrieve a task by ID with full state history | | tasks/list | List tasks for the project with optional filters | | tasks/cancel | Cancel a running task | | tasks/resubscribe | Resume streaming updates for a running task |

Continue a conversation by passing contextId:

{
  "jsonrpc": "2.0",
  "method": "message/send",
  "id": "2",
  "params": {
    "contextId": "context-uuid-from-previous-response",
    "message": {
      "role": "user",
      "parts": [{"kind": "text", "text": "Tell me more about the first item"}]
    }
  }
}

Task Management

Tasks track the lifecycle of message requests. States: submitted, working, input-required, completed, canceled, failed.

{
  "jsonrpc": "2.0",
  "method": "tasks/get",
  "id": "1",
  "params": {"id": "task-uuid", "historyLength": 10}
}

A2A REST Endpoints

| Method | Endpoint | Description | |--------|----------|-------------| | GET | /.well-known/agent.json | Root agent card | | GET | /projects | List all projects with agent URLs | | POST | /projects | Create a new project | | GET | /projects/:id | Get complete project data | | GET | /projects/:id/.well-known/agent.json | Project agent card | | GET | /projects/:id/contexts | List conversations | | GET | /projects/:id/contexts/:ctxId/messages | Get message history | | GET | /projects/:id/action-items | List action items (supports filters) | | POST | /projects/:id/action-items | Create action items | | PATCH | /projects/:id/action-items/:itemId | Update an action item | | POST | /projects/:id/a2a | JSON-RPC messaging | | GET | /workspaces | List all workspaces with active workspace ID | | POST | /workspaces/:id/activate | Switch to a different workspace |

Optional Auth: Set A2A_HTTP_AUTH_TOKEN env var before starting TensorPM to enable token validation.

Available MCP Tools

| Tool | Description | |------|-------------| | list_projects | List all projects with names and IDs | | create_project | Create a new project (basic, fromPrompt, or fromFile mode) | | get_project | Get complete project data (read-only) | | list_action_items | Query and filter action items | | submit_action_items | Create new action items | | update_action_items | Update existing action items | | propose_updates | Submit project updates for human review | | set_api_key | Set AI provider API key (openai, anthropic, google, mistral) | | list_workspaces | List all workspaces (local + cloud) with active workspace ID | | set_active_workspace | Switch to a different workspace |

Note: MCP tools provide direct access to action items. Core project context (profile, budget, people, categories) can only be modified by the TensorPM project manager agent — use A2A message/send to request changes.

Tool parameters: Use the MCP tool schemas for detailed parameter information.

Action Item Fields

...