A comprehensive framework for deploying an autonomous AI companion on Feishu featuring long-term memory, voice messaging, and image generation.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawgirl
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install clawgirl using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
This skill provides a complete blueprint for creating a sophisticated AI companion within the Openclaw Skills ecosystem. By leveraging the Openclaw multi-agent framework, developers can craft agents that possess a unique soul (personality), a rhythmic heartbeat (automated behaviors), and the ability to interact via text, high-quality voice messages, and realistic selfies.
The framework is specifically designed to overcome common limitations in automated messaging, such as session isolation and media format requirements, providing a battle-tested architecture for persistent AI personas. Using this approach within Openclaw Skills ensures that your agent remains context-aware and human-like in its interactions.
To deploy this companion within your Openclaw Skills environment, follow these steps:
brew install ffmpeg jq
pip install edge-tts fal-client
openclaw.json with the necessary environment variables:FAL_KEY: For image and high-quality TTS generation.FEISHU_APP_ID & FEISHU_APP_SECRET: For messaging API access.SOUL.md, HEARTBEAT.md, and the memory/ directory.openclaw cron add --name "companion-heartbeat" --every 15m --session isolated --message "Follow HEARTBEAT.md logic and send messages via send.sh" {agent_id}
The skill organizes its internal logic and state using the following file structure within the agent workspace:
| File/Folder | Purpose |
|---|---|
SOUL.md |
Defines identity, persona, linguistic style, and anti-AI constraints. |
HEARTBEAT.md |
Maps specific times and contexts to agent behaviors and actions. |
TOOLS.md |
Technical documentation for the agent to understand how to use bash scripts. |
memory/ |
Directory containing YYYY-MM-DD.md logs for long-term context retention. |
scripts/ |
Contains send.sh, voice.sh, and selfie.sh for platform interaction. |
Loading
A comprehensive framework for deploying lifelike AI agents with personality, persistent memory, and multi-modal communication capabilities.

A comprehensive framework and configuration guide designed to prevent AI agents from freezing during large-scale batch operations by managing session transcript bloat.

A production-grade workspace enhancement for OpenClaw agents that implements enterprise-level artifact management, security protocols, and memory optimization.

A compatibility layer designed to transition users from legacy monolithic leak skills to hardened, specialized Openclaw Skills.

An automated browser-based skill designed to handle the entire lifecycle of Feishu open platform application configuration for AI agents.

A production-grade blueprint for orchestrating a team of AI agents through Feishu using a foreman-worker model and file-driven task queues.








































