iamdibo365

Argus Agent — AI skill for Claude Code

AI community

How to install Argus Agent

This entry records only its repository, not the path inside it, so there is no exact command to give. Open iamdibo365/argus-agent and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Argus Agent does

Argus is a production-style single AI agent that lives in your Slack workspace and can search and read your Google Drive, summarize channel discussions, and post messages — built with **LangChain / LangGraph**, **OpenAI** (or **AWS Bedrock** via one env var), **LangSmith** tracing, and deployed as a container on **AWS ECS Fargate**.

Alternatives in AI

Manifest — Real-time cost observability for OpenClaw agents — track tokens, costs, messages, and model usage 4.1k ★
Wechat Acp — Bridge WeChat chat messages to any ACP-compatible AI agent (Claude, Codex, Copilot, Qwen, Gemini, OpenCode and 488 ★
Academic Writing Skills — AI-powered post-writing toolkit for academic papers — format validation, grammar/style polishing, de-AI editin 430 ★

README

Argus — AI Agent for Slack + Google Drive on AWS

Roadmap: Argus v2 evolves into a multi-agent team (researcher, writer, reviewer) orchestrated with LangGraph — this repo is the single-agent foundation.

Architecture

Slack (Socket Mode websocket)
        │
        ▼
┌─────────────────────────────┐
│  ECS Fargate container       │
│  ┌────────────────────────┐  │        ┌──────────────┐
│  │ Slack Bolt app          │  │◄──────►│  LangSmith   │ (tracing)
│  │   └─ LangGraph ReAct    │  │        └──────────────┘
│  │      agent              │  │
│  │   Tools:                │  │        ┌──────────────┐
│  │   • search_drive        │──┼───────►│ Google Drive │
│  │   • read_drive_file     │  │        └──────────────┘
│  │   • send_slack_message  │  │        ┌──────────────┐
│  │   • get_channel_history │──┼───────►│  Slack API   │
│  └────────────────────────┘  │        └──────────────┘
│  LLM: OpenAI GPT-4o  ⇄  AWS Bedrock (Claude) — env var swap
└─────────────────────────────┘
   Secrets: SSM Parameter Store · Logs: CloudWatch

**Why Socket Mode?** The container connects *out* to Slack over a websocket, so there is no public endpoint, no API Gateway, and no ALB — smaller attack surface, simpler infra, and no Slack 3-second-ack gymnastics.

Project structure

argus-agent/
├── README.md
├── .env.example
├── .gitignore
├── requirements.txt
├── Dockerfile
├── src/argus/
│   ├── __init__.py
│   ├── config.py           # pydantic-settings config
│   ├── llm.py              # OpenAI ⇄ Bedrock factory
│   ├── agent.py            # LangGraph ReAct agent + per-thread memory
│   ├── app.py              # Slack Bolt (Socket M