A sophisticated tool designed to distill real-world colleagues into interactive AI personas by analyzing Feishu, DingTalk, and email data.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install create-colleague
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 create-colleague using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The Create Colleague skill is a powerful framework for Openclaw Skills that allows developers to clone the professional expertise and personality of a teammate into a functional AI agent. By analyzing communication patterns in Feishu or DingTalk and reviewing technical output, it creates a digital twin capable of acting with the same technical rigor and communication style as the original person.
This skill is particularly valuable for preserving institutional knowledge and simulating peer reviews. It utilizes a dual-track analysis system to separate technical capabilities from persona traits, ensuring that the resulting AI agent not only knows how to solve a problem but also interacts in a way that feels authentic to the person it represents.
To get started with this component of Openclaw Skills, ensure you have the necessary environment variables and dependencies installed. Run the following setup command for the primary collection tool:
# Initialize the Feishu auto-collector
python3 ${CLAUDE_SKILL_DIR}/tools/feishu_auto_collector.py --setup
# Or initialize for DingTalk
python3 ${CLAUDE_SKILL_DIR}/tools/dingtalk_auto_collector.py --setup
You will also need to configure your Feishu App ID and Secret if you plan to use automated private chat collection.
The skill organizes distilled colleague data into a structured directory under ./colleagues/{slug}/:
| File | Description |
|---|---|
SKILL.md |
The primary entry point for the AI agent, containing PART A (Work) and PART B (Persona). |
work.md |
Detailed technical standards, responsible systems, and workflow preferences. |
persona.md |
The 5-layer behavioral and communication framework of the colleague. |
meta.json |
Metadata including versioning, timestamps, and imported knowledge sources. |
knowledge/ |
A subdirectory containing raw extracted text from messages, docs, and emails. |
Loading
A specialized financial intelligence skill for analyzing Fannie Mae's role in the US housing finance system and mortgage-backed securities market.

A comprehensive business intelligence module analyzing GoPro's growth, market challenges, and pivot to a subscription-based ecosystem.

An analytical research skill focused on the business strategy, historical evolution, and market moat of the world's leading anti-dandruff brand.

A comprehensive strategic analysis of how Grey Goose broke the Russian vodka monopoly through French craftsmanship and ultra-premium positioning.

An intelligent utility for auto-optimizing AI prompts through multi-dimensional analysis, A/B testing, and cost-reduction techniques.

A powerful command-line utility for displaying, formatting, and calculating date and time values within AI agent workflows.








































