ClawCode plugin
A Claude Code plugin that gives the agent a name, a personality and memory across sessions, plus chat channels.
- 64 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The agent reads and writes messages in WhatsApp, Telegram, Discord and Slack and runs around the clock.
- Webhooks pass outside events to the agent, so request text can be spoofed through them.
Install
- Works in
- Services
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/clawcodeDetects 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.
- The agent reads and writes messages in WhatsApp, Telegram, Discord and Slack and runs around the clock.
- Webhooks pass outside events to the agent, so request text can be spoofed through them.
This is third-party code. Review the repository files before installing.
What it does
ClawCode turns Claude Code sessions into a persistent agent. The SessionStart, PreCompact, Stop and SessionEnd hooks load personality and memory files when work begins, save facts before context compression and write a session summary. Memory runs a bilingual recall over notes, so a question in Spanish finds an entry written in English. At night the agent consolidates notes into long-term memory in three phases. WhatsApp, Telegram, Discord, iMessage and Slack connect through MCP plugins, with a browser chat and webhooks for outside systems. Each agent folder is self-contained: personality, memory and settings live inside it. A launchd or systemd service enables round-the-clock operation.
Who it is for. Claude Code users who want a personal assistant with a name, memory of past conversations and replies in a familiar messenger, without a separate runtime.
Good fit when
- Every new Claude Code session starts from zero, and you need an agent that remembers yesterday's decisions.
- The agent should reply to you in Telegram or WhatsApp while you are away from the computer.
- You are moving from OpenClaw and want to carry personality, memory and schedules over with `/agent:import`.
- Reminders and the nightly memory digest must run around the clock as a service.
- External systems such as GitHub Actions should send the agent events over HTTP.
Not a fit when
- You need a one-off coding helper without personality or memory: plain Claude Code is enough.
- Windows without WSL2: Claude Code has to run inside WSL2, native PowerShell does not work.
- Other people's chats must stay closed to the agent and you do not plan to configure the channel scope mode.
Example request
/agent:createLimitations
Node.js 18 or newer is required. On Windows, Claude Code runs from WSL2. Schedules run only while Claude Code is open; for round-the-clock operation install the service with `/agent:service install`. Messaging channels connect through separate MCP plugins, each with its own tokens and platform rules. The hooks add text from the personality and memory files to the context, so surplus notes cost tokens. The project is independent and not affiliated with Anthropic: updates arrive through the community marketplace or the author's own one, and fresh fixes appear in the shared catalog with up to a day of delay.
How to disable. Remove the plugin with `/plugin uninstall` inside Claude Code. The agent folder with personality and memory stays on disk; delete it by hand.
FAQ
How do I install ClawCode in Claude Code?
Run `/plugin marketplace add anthropics/claude-plugins-community`, then `/plugin install clawcode@claude-community` and `/reload-plugins`. After that `/agent:create` starts the introduction where the agent picks a name, personality and emoji.
How does the plugin differ from a standalone runtime such as OpenClaw?
A runtime of that kind sits on the model API with its own sessions and billing. ClawCode extends Claude Code itself with hooks, skills and MCP, so there is no separate billing and the Claude Code subscription stays the base.
How do I keep the agent on around the clock?
Run `/agent:service install`. A launchd service on macOS or systemd on Linux restarts the agent after a reboot and enables the nightly memory digest and scheduled checks.
How do I limit the agent's access to chats?
`/agent:scope wizard` configures channel scope: the agent sees all chats as owner, sees nothing as guest, or runs in auto mode. Until you enable it, behavior does not change.
Can I move an agent over from OpenClaw?
Yes, `/agent:import` carries over personality, memory, skills and schedules. Skills and schedules are sorted into three classes, green, yellow and red, and skipped items go to a backlog with recovery notes.
Discussion
No questions or reviews yet.
Related
All Telegram toolsA compact OpenClaw alternative: a personal assistant on Claude Agent SDK with agents in isolated containers and messengers added via skills
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
Claude-to-IM
claude-to-im
A bridge between Claude Code or Codex and messengers: talk to the agent from Telegram, Discord, Feishu, QQ or WeChat
Self-hosted monitoring and remote approvals for Claude Code and Codex via Telegram, Feishu or a web terminal
Connects Claude Code, Cursor, Codex or OpenCode to Telegram, Slack and Discord: teammates message a bot, an agent replies
About a hundred skills for Claude Code: meetings, research, publishing, TDD, Telegram, email and personal analytics
README in short
The author builds a persistent agent inside Claude Code rather than on top of its API. Personality comes from SOUL.md and IDENTITY.md, memory sits in `memory/` as daily notes. An OpenClaw import sorts skills and schedules into three tiers.