guru-mcp

What This Skill Does

The guru-mcp skill integrates OpenClaw directly with your Guru knowledge base, allowing for seamless AI-powered document retrieval, card management, and cross-platform information synthesis. By leveraging the Model Context Protocol (MCP), this skill provides an intelligent bridge between your existing workflows and your organizational brain. Whether you need to query documentation from connected sources like Slack, Confluence, or Google Drive, or you need to maintain your internal wiki by creating drafts or updating existing cards, this skill streamlines the entire knowledge lifecycle within your agentic environment.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/pvoo/guru-mcp. Once installed, ensure you have your API token generated from the Guru Admin dashboard. Create a .env file at ~/.clawdbot/.env and add the GURU_API_TOKEN in the format [email protected]:your-api-token. Finally, update your config/mcporter.json to register the server, and verify the connection using mcporter list guru. Always consult the provided documentation if you encounter authentication errors.

Use Cases

• Unified Knowledge Search: Quickly retrieve answers from across fragmented platforms like Jira, SharePoint, and Slack without switching tabs.
• Documentation Maintenance: Automate the creation of draft cards after a meeting or update outdated process guides directly via natural language commands.
• Support Efficiency: Empower agents to generate accurate, source-cited responses for internal team members by querying the internal Knowledge Base.

Example Prompts

1. "Check our internal Guru knowledge base for the latest onboarding process documentation and summarize it for me."
2. "Search for all cards related to 'Q4 project roadmap' and draft a new update for the team based on these findings."
3. "How do I request time off according to the HR documentation stored in Guru? Provide the steps clearly."

Tips & Limitations

• Permission Integrity: This skill respects all existing Guru permissions, meaning if you don't have access to a collection in Guru, the AI will not be able to retrieve content from it.
• Initialization: Always run guru_list_knowledge_agents at the start of your session to ensure you are targeting the correct workspace ID.
• Data Structure: When updating cards, remember that content must be provided in HTML format. It is best practice to always call guru_get_card_by_id first to retrieve the current HTML structure before attempting an update to avoid data loss.