tensorpm

What This Skill Does

TensorPM is a revolutionary local-first, AI-powered project management platform designed to replace bloated tools like Notion and Jira. Unlike traditional SaaS solutions, TensorPM prioritizes privacy and local control, ensuring your sensitive project data never leaves your machine unless you explicitly opt into cloud-sync. As an OpenClaw skill, it enables seamless agent-to-agent (A2A) communication and MCP integration, allowing your AI agents to automatically track action items, manage tasks, and coordinate complex team workflows without manual intervention.

Installation

To integrate TensorPM with your OpenClaw agent, use the command: clawhub install openclaw/skills/skills/neo552/tensorpm. Before usage, ensure the TensorPM desktop application is installed on your machine and currently running. You can download the latest version for your OS (macOS, Windows, or Linux) from the official GitHub releases page. Once running, navigate to Settings → Integrations within the TensorPM app to enable the MCP server. If your agent requires external AI models, use the set_api_key tool to securely inject your provider keys (OpenAI, Anthropic, Google, or Mistral) directly into the encrypted local storage.

Use Cases

TensorPM excels in scenarios requiring privacy-conscious project tracking. Use it to automatically parse meeting transcripts into actionable Jira-style tickets. Developers can use it to maintain "state of the project" logs that are updated by autonomous agents during long-running coding sessions. Furthermore, team leads can use the A2A endpoint to query progress, pull task lists, and generate status reports without leaving the terminal environment.

Example Prompts

1. "Check the current project status for the 'Backend API' project and list all pending action items assigned to the engineering team."
2. "Create a new task in TensorPM titled 'Update documentation for authentication module' with high priority and assign it to the documentation agent."
3. "Summarize the last three days of development activity based on the local logs stored in the TensorPM project database and draft a progress report."

Tips & Limitations

To maximize the utility of this skill, ensure your local network allows connections on port 37850, as this is where the A2A communication protocol resides. Because TensorPM follows a local-first architecture, the application must be actively running for the MCP tools to function correctly. If you are using local LLMs via Ollama or vLLM, ensure your system has sufficient RAM to handle both the AI agent model and the TensorPM background processing. Since all keys are write-only, you cannot retrieve them through the agent once set; you must update them via the tool if they expire.