add-mnemon

Installs mnemon persistent graph-based memory into NanoClaw agent containers via Dockerfile and entrypoint changes.

2|Updated Jan 14, 2026
One-click install
npx skills add https://github.com/RADHA0-max/selfhealbackend --skill add-mnemon-radha0-max
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: add-mnemon
Source: https://github.com/RADHA0-max/selfhealbackend/tree/main/nanoclaw/.claude/skills/add-mnemon
Command: npx skills add https://github.com/RADHA0-max/selfhealbackend --skill add-mnemon-radha0-max

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires vitest.

What problem does it solve? NanoClaw agents lose all conversational context when containers restart, forcing users to restate prior information in every new session. This Skill wires the mnemon memory engine into the agent container image so agents recall past context before responding and store new insights after each turn. ## Core Features & Use Cases - Container Image Integration: Adds a Dockerfile layer that downloads the mnemon release binary and sets MNEMON_DATA_DIR into the persistent .claude mount. - Automatic Hook Registration: Runs idempotent mnemon setup in the container entrypoint to register Claude Code hooks on every start. - Structural Test Guards: Ships vitest tests that fail if the Dockerfile install layer or entrypoint wiring is dropped during upgrades. - Use Case: Apply this Skill to a NanoClaw deployment, rebuild the image, and agents will automatically surface relevant memory from earlier conversations in new sessions without restating context. ## Quick Start Ask the agent to apply the add-mnemon skill to install persistent memory into the NanoClaw agent container and rebuild the image.

Frequently Asked Questions about add-mnemon

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

FAQPage Schema
How do I add persistent memory to NanoClaw agents?

Apply the add-mnemon skill, which inserts a mnemon binary install layer into container/Dockerfile and a mnemon setup line into container/entrypoint.sh. Rebuild the image with ./container/build.sh and restart the service to activate memory hooks.

How does mnemon memory persist across container restarts?

MNEMON_DATA_DIR points to /home/node/.claude/mnemon inside the container, which maps to the per-agent-group .claude/ mount on the host. Because data is written to the mounted volume, memory survives container restarts and rebuilds.

Does mnemon work with non-Claude providers like opencode?

No. mnemon hooks fire only under --target claude-code, so groups whose container.json sets a different provider such as opencode never invoke the claude CLI and the hooks do not run. Check the provider key in groups/*/container.json before applying.

Why does the container show mnemon: command not found?

This error means the image was not rebuilt after adding the Dockerfile install layer. Run ./container/build.sh to rebuild the image, then restart the service so the new binary is available in the container.

How do I remove mnemon memory from an agent?

Follow REMOVE.md: delete the mnemon block from the Dockerfile, remove the setup line from entrypoint.sh, delete the test files, rebuild, and restart. Optionally delete the mnemon/ subdirectory from the host .claude mount to erase stored memory.