Hivemind plugin
Shared cloud memory for a team's agents: stores session traces and turns repeated patterns into skills
- 1,641 ★ on the repository
- Last commit
- Apache-2.0
Broad access
Why this level
- Sends prompts, tool calls and responses from every session to an external cloud
- Installs hooks into every detected assistant
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/hivemindDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Sends prompts, tool calls and responses from every session to an external cloud
- Installs hooks into every detected assistant
This is third-party code. Review the repository files before installing.
What it does
Hivemind records every agent session's prompts, tool calls and responses as structured traces in Deeplake. A background worker mines traces for repeated patterns and codifies them into SKILL.md files available to every agent in the workspace. Search over traces and skills is hybrid lexical and semantic, and a wiki page summarizing each session is generated at the end. It connects to Claude Code as a plugin, to Codex and Cursor through hooks, and to OpenClaw, Hermes and pi through extensions and MCP.
Who it is for. For engineering teams that want one agent's findings to become available to teammates' agents.
Good fit when
- A team of several people uses Claude Code, Codex or Cursor on shared work
- Past session solutions should be recalled in new sessions automatically
- You want skills distilled from real work rather than written by hand
Not a fit when
- Code and prompts must not go to an external cloud
- You work solo and the agent's local memory is enough
Example request
Search team memory for how we handled the last database migration with column renamesLimitations
Requires a Deeplake account and sign-in via browser or token. Session traces, including prompts and tool output, are stored in Deeplake's cloud or your own bucket with BYOC. The npm install needs Node.js 22. Claude Cowork support is alpha. Availability from Russia without a VPN is unverified.
How to disable. Run hivemind uninstall for all assistants or hivemind codex uninstall for one. Capture for a session is disabled with HIVEMIND_CAPTURE=false.
FAQ
Can data stay in my own storage?
Yes, BYOC supports your own GCS, Azure, S3 or on-prem bucket.
How to install without a browser, e.g. in CI?
Pass a token: HIVEMIND_TOKEN=<token> hivemind install or hivemind install --token <token>.
Discussion
No questions or reviews yet.
Related
PluginsLocal persistent memory for coding agents: hooks record every session, and Claude Code, Codex or Cursor recall past decisions through 54 MCP tools
A long-term memory plugin that keeps decisions, preferences and task state across sessions and injects only what a query needs
Long-term memory for AI agents: a skill and the mindmemos command store facts from conversations and recall them in later sessions
Open infrastructure for agents that learn from experience: session traces, a knowledge base and skill packs via CLI, MCP and API
Shared local memory for AI agents: Claude Code, Codex, Cursor and others find the agreements of past sessions in it and record new ones
Knowledge-graph memory for agents: Claude Code and Codex plugins, an MCP server and a Python library
README in short
The README describes the loop of capturing traces, codifying skills and propagating them across a team's agents, and reports the authors' LoCoMo benchmark results. Install is a one-line script, npm, or a Claude Code plugin, with a table of supported agents and integration methods. A section covers tenant isolation, encryption and storage. Apache-2.0 licensed.