Skills cc-connect
git clone https://github.com/TerminalSkills/skills
T=$(mktemp -d) && git clone --depth=1 https://github.com/TerminalSkills/skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/cc-connect" ~/.claude/skills/terminalskills-skills-cc-connect && rm -rf "$T"
skills/cc-connect/SKILL.md- global npm install
CC-Connect
Overview
CC-Connect bridges AI coding agents running on your local machine to the messaging platforms your team already uses. Code review, research, automation, data analysis — anything an AI agent can do, now accessible from your phone, tablet, or any device with a chat app.
Architecture: Your local AI agent <-> CC-Connect bridge <-> Messaging platform (Slack/Telegram/Discord/etc.)
Send a message in Slack, CC-Connect routes it to your local Claude Code instance, the agent does the work, and the response comes back to your chat.
Instructions
Installation
npm install -g cc-connect
Configuration
Create a
cc-connect.yaml in your project:
agent: type: claude-code # or: codex, gemini, cursor workdir: /path/to/your/project platform: type: telegram # or: slack, discord, feishu, dingtalk token: "your-bot-token"
Platform Setup
Telegram: Create a bot via @BotFather, get the bot token, and add it to your config.
Slack: Create a Slack App at api.slack.com/apps, enable Socket Mode and Event Subscriptions, add Bot Token Scopes (
chat:write, app_mentions:read, messages.im), and install to your workspace.
Discord: Create an application at discord.com/developers, create a bot, enable Message Content Intent, and invite the bot to your server.
Starting the Bridge
cc-connect init # Interactive wizard for platform credentials cc-connect start # Start routing messages
Session Management
session: timeout: 30m max_concurrent: 3 continue: true auto_compress: true
Multi-Agent Routing
Route different commands to different agents:
agents: code-review: type: claude-code workdir: /path/to/project trigger: "!review" research: type: gemini trigger: "!research"
Access Control
access: allowed_users: ["U123", "U456"] allowed_channels: ["C789"] admin_users: ["U123"]
Examples
Example 1: Team Code Review via Slack
A team sets up CC-Connect to allow engineers to request code reviews from Slack:
# cc-connect.yaml agent: type: claude-code workdir: /home/dev/acme-api platform: type: slack app_token: "xapp-1-A07QX4R..." bot_token: "xoxb-8234567890-..." channels: ["#code-review"] session: timeout: 10m auto_compress: true access: allowed_channels: ["#code-review"] allowed_users: ["U0381KDLS", "U0492JFMA"]
In Slack
#code-review, an engineer types: @agent Review the auth module for SQL injection risks. Claude Code analyzes the code and responds in the thread with findings.
Example 2: Scheduled Daily Reports via Telegram
A solo developer configures CC-Connect with cron jobs for automated daily standup reports:
agent: type: claude-code workdir: /home/dev/saas-app platform: type: telegram token: "7284619035:AAF-kLm9xPqR..." allowed_users: ["198274563"] cron: - schedule: "0 9 * * 1-5" command: "Summarize yesterday's git commits and open PRs, highlight blockers" platform: telegram timeout: 5m fresh_session: true
Every weekday at 9am, the agent generates a summary of recent activity and sends it to the developer's Telegram chat.
Guidelines
- Start with one messaging platform and get it working before expanding to others
- Always set
in production to restrict accessallowed_users - Use threads in Slack/Discord to keep conversations organized
- Set
to prevent runaway agent sessions consuming resourcessession.timeout - Enable
for long conversations to prevent context overflowauto_compress - Use
for cron jobs to avoid inherited context from previous runsfresh_session: true - Verify your setup with
if messages are not routingcc-connect status - See the GitHub Repository for full documentation