rune

Store and retrieve encrypted organizational decisions across agents using Rune's FHE-based memory.

19|8|Updated Feb 2, 2026
One-click install
npx skills add https://github.com/CryptoLabInc/rune --skill rune
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: rune
Source: https://github.com/CryptoLabInc/rune/tree/main
Command: npx skills add https://github.com/CryptoLabInc/rune --skill rune

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyenvector, fastmcp, psutil, prometheus-client, httpx, pydantic, fastembed, langdetect, langchain-text-splitters, pypdf, and includes scripts (resource) components.

What problem does it solve?

Rune provides a secure, end-to-end encrypted organizational memory workflow that enables teams to capture, store, and retrieve institutional knowledge across multiple agents and sessions, without exposing plaintext data to external services.

Core Features & Use Cases

  • Cross-agent memory and recall: works with Claude Code, Codex CLI, Gemini CLI, and MCP-compatible agents to surface past decisions.
  • Activation-aware runtime: validates local bootstrap, vault/envector configuration, and maintains active/dormant states with /rune commands.
  • Zero-knowledge privacy: data is encrypted client-side and decrypted only by authorized vault members, with encrypted indexing on enVector Cloud.
  • Use case: record architectural decisions and retrieval trails to guide future planning and audits.

Quick Start

Configure credentials and enable Rune with a single command: /rune:configure to set up credentials, start MCP servers, and activate the plugin.

Frequently Asked Questions about rune

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I store and retrieve encrypted organizational memory across multiple AI agents?

You use Rune to store and retrieve encrypted organizational memory across AI agents via FHE-based memory. It coordinates cross-agent capture and recall using /rune commands and Vault/envector credentials for secure collaboration.

Does encrypted memory for AI teams work with Claude Code, Codex CLI, and Gemini CLI?

Yes, encrypted memory for AI teams works with Claude Code, Codex CLI, Gemini CLI, and any MCP-compatible agent. Rune provides cross-agent memory and recall to surface past decisions across these platforms.

How do I configure credentials and activate encrypted memory for my AI agents?

To configure encrypted memory credentials, run the /rune:configure command. This sets up your Vault/envector credentials, starts the MCP servers, and activates the plugin to establish an active runtime state.

What is zero-knowledge encrypted memory and how does it protect organizational decisions?

Zero-knowledge encrypted memory protects organizational decisions by encrypting data client-side and decrypting it only for authorized vault members. Rune uses encrypted indexing on enVector Cloud to ensure plaintext is never exposed to external services.

Can I record architectural decisions and retrieval trails to guide future planning and audits?

Yes, you can record architectural decisions and retrieval trails to guide future planning and audits. Rune captures this institutional knowledge across sessions, maintaining an encrypted organizational memory workflow for AI teams.

What happens if the local bootstrap or vault configuration fails during AI agent memory activation?

Rune features an activation-aware runtime that validates local bootstrap and vault/envector configuration before maintaining active or dormant states. This resilient activation lifecycle ensures secure context access even if configuration validation encounters issues.