ClawHub
Code Pluginsource linked

Hivemindv0.7.141

Hivemind — cloud-backed persistent shared memory for AI agents, powered by DeepLake

hivemind·runtime hivemind·by @kaghni
Community code plugin. Review compatibility and verification before install.
openclaw plugins install clawhub:hivemind
Latest release: v0.7.141Download zip

Compatibility

Built With Open Claw Version
2026.3.22
Min Gateway Version
>=2026.3.22
Plugin Api Range
>=1.0.0
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
high confidence
!
Credentials
The plugin reads and uploads user and assistant messages, indexes repository structure, uses a long-lived Deeplake token, and runs external agent CLIs over session-derived prompts.
ℹ
Install Mechanism
The package installs as an OpenClaw plugin and setup commands modify OpenClaw config with backups; this is documented and command-triggered, but it changes tool allowlists and auto-update settings.
!
Instruction Scope
The runtime injects broad memory instructions and the embedded documentation says OpenClaw does not run sessions to mine, while the code spawns a skillify worker after capture, creating a disclosure mismatch.
!
Persistence & Privilege
It launches detached background graph, pull, and skill-mining workers; graph builds can push snapshots to Deeplake, and skillify can write global Claude skills and remote skill rows.
!
Purpose & Capability
Cloud-backed memory, auto-capture, team sharing, and code graph indexing are mostly coherent with the stated purpose and are disclosed, but automatic skill-mining from prior sessions materially expands the plugin beyond shared memory.
Scan Findings in Context
[SDI-2] expected: Automatic conversation capture to Deeplake is high-impact but directly matches the cloud memory purpose and is disclosed in the README and manifest.
[SQP-3] expected: Org-wide memory sharing is sensitive, but the README explicitly says memory is shared with teammates in the Deeplake org.
[PE3] expected: The local long-lived credential is disclosed and needed for the Deeplake integration, though users should treat it as sensitive.
[SQP-2] expected: Config changes for allowlisting and auto-update are implemented by explicit commands and backups, so this is a notable permission change rather than hidden install-time behavior.
[SDI-2] unexpected: The code automatically spawns a skillify worker after successful capture; that worker processes stored sessions and can create or merge skills, which is not clearly covered by the primary memory-plugin framing.
[SDI-4] unexpected: The embedded instructions say OpenClaw itself does not mine sessions, but the runtime does spawn the skillify worker after session capture.
[SQP-2] unexpected: The skillify worker invokes external agent CLIs with bypass or yolo-style flags and relies on prompt instructions to limit writes, which is under-scoped for automatic background execution.
[RA2] unexpected: Detached graph and post-commit execution are purpose-adjacent, but the build path can also push cloud snapshots and trigger docs refresh when enabled, so users need clearer control.
What to consider before installing
Review carefully before installing. This plugin intentionally sends conversations to Deeplake and shares them across the selected org. The main concern is the automatic background skill-mining path: it can run external agent CLIs with broad permissions and write skills that affect future agent behavior. Install only if you want org-wide shared memory plus these automations, and disable capture, graph push/pull, docs auto-sync, or skillify behavior where your environment requires tighter control.

Verification

Tier
source linked
Scope
artifact only
Summary
Validated package structure and linked the release to source metadata.
Source
github.com/activeloopai/hivemind
Commit
5928e1f65f1f
Tag
5928e1f65f1fe81bdbeb8b371b26421e72734d97
Provenance
No
Scan status
suspicious

Tags

latest
0.7.141
test
0.0.1-test

Hivemind

Cloud-backed shared memory for AI agents. Install once, memory persists across sessions, machines, and channels — and is shared with every teammate in your Deeplake org.

Powered by Deeplake.

Install

openclaw plugins install clawhub:hivemind

Then in chat:

/hivemind_login

Click the auth link, sign in, send another message. That's it.

What it does

  • Auto-recall — before each agent turn, relevant memories surface automatically via keyword search.
  • Auto-capture — after each turn, the conversation is stored to your Deeplake workspace.
  • Code graph — AST-derived map of your repo (TS/JS/Python) auto-rebuilds after each turn; query via hivemind_graph_search and hivemind_graph_neighborhood tools.
  • Cross-platform — same memory accessible from Claude Code, Codex CLI, and OpenClaw plugins.
  • Team-wide — every user in your Deeplake org shares the same memory.

Commands

CommandWhat it does
/hivemind_loginSign in via device flow
/hivemind_setupAdd hivemind to OpenClaw's tool allowlist (one-time, after install)
/hivemind_captureToggle conversation capture on/off
/hivemind_whoamiShow current org and workspace
/hivemind_orgsList organizations
/hivemind_switch_org <name>Switch organization
/hivemind_workspacesList workspaces
/hivemind_switch_workspace <id>Switch workspace
/hivemind_versionShow installed version and check ClawHub for a newer one
/hivemind_updateShow how to install the latest version
/hivemind_autoupdate [on|off]Toggle the agent-facing update nudge (on by default)

You can also just ask the agent naturally — "switch org to activeloop", "list my orgs", "invite alice@example.com as admin", etc.

Privacy & data

  • What's captured: every user message and assistant reply, sent to api.deeplake.ai.
  • Where credentials live: a long-lived API token at ~/.deeplake/credentials.json (file permissions 0600).
  • Where it sends data: api.deeplake.ai (memory storage) and clawhub.ai (version check on session start and via /hivemind_version).
  • How to pause: run /hivemind_capture to stop capture; run it again to resume.
  • How to fully sign out: delete ~/.deeplake/credentials.json and revoke the token in the Deeplake dashboard.

OpenClaw config changes

The plugin modifies ~/.openclaw/openclaw.json in two places, both triggered by explicit user commands and both with timestamped backups:

  • /hivemind_setup appends "hivemind" to tools.alsoAllow so OpenClaw admits the plugin's agent tools. OpenClaw's default coding profile only exposes core tools (read/write/exec/etc.) to agents; plugin-registered tools are filtered out unless explicitly allowed.
  • /hivemind_autoupdate [on|off] sets plugins.entries.hivemind.config.autoUpdate. When on, the plugin adds a short line to the system prompt when a newer version is available on ClawHub; the actual install runs through the agent's existing exec tool or via openclaw plugins update hivemind in a terminal.

The plugin does not replace the built-in memory plugin. It runs alongside memory-core via lifecycle hooks, so memory-core's dreaming cron and other memory-slot jobs keep working.

Troubleshooting

Hivemind feels slow or makes tools hang. Check agents.defaults.model in ~/.openclaw/openclaw.json. Memory-heavy workflows issue many small tool calls; a large reasoning model feels sluggish. Recommended default is anthropic/claude-haiku-4-5-20251001.

Model switch rejected as "not allowed". OpenClaw's allowlist wants <provider>/<exact-dated-id>. Use anthropic/claude-haiku-4-5-20251001 or anthropic/claude-sonnet-4-6. Bare IDs and -latest suffixes are rejected.

openclaw model <id> fails with "plugins.allow excludes model". The CLI is disabled by default. Edit ~/.openclaw/openclaw.json under agents.defaults.model and restart the gateway: systemctl --user restart openclaw-gateway.service.

Telegram-triggered sudo npm i -g openclaw@latest fails with "elevated is not available". tools.elevated.allowFrom.telegram isn't set. Run the upgrade in a local shell instead.

Sharing memory with teammates

Invite teammates to your Deeplake org:

invite alice@example.com as admin

Their agents will see your memory; your agents will see theirs.

Source

github.com/activeloopai/hivemind