stablyai/agent-slack

Slack automation CLI for AI agents

What it solves

agent-slack is a CLI tool designed specifically for AI agents to interact with Slack. It solves the problem of providing agents with a token-efficient, easy-to-authenticate, and safe way to read, write, and manage Slack communications without requiring complex API setup or Python dependencies.

How it works

The tool is built with TypeScript and Bun, providing a set of commands for agents to perform Slack actions. It focuses on three core principles:

  • Token Efficiency: It outputs aggressively pruned JSON (removing null and empty fields) to reduce the cost and token usage for LLMs.
  • Zero-Config Auth: It automatically extracts authentication tokens from Slack Desktop or supported browsers (Chrome, Brave, Firefox) on macOS and Windows, removing the manual token management burden.
  • Human-in-the-Loop: It includes a "Safe Mode" that redirects send, edit, and delete operations to a browser-based editor or drafts, ensuring no message is posted without human review.

Who it’s for

Developers building AI agents (using frameworks like Claude Code, Codex, or Cursor) who need their agents to be able to read and write to Slack channels, manage threads, and search for information within a workspace.

Highlights

  • Comprehensive Slack Control: Supports reading messages, listing threads, searching messages/files, creating channels, and managing Slack canvases as Markdown.
  • Agent-Friendly Artifacts: Automatically downloads attached files (snippets, images) to local paths so agents can read them directly.
  • Smarter Writing: Converts bullet/numbered lists to native Slack rich text and supports Block Kit for structured layouts.
  • Native Drafts: Can create and update Slack-native drafts that appear in the user's Slack client across devices.
  • Safe Mode: An enforceable tool-level guarantee that prevents agents from posting directly to Slack, redirecting them to drafts or a browser editor instead.

Related

  • Project
    clacky-ai/openclacky

    An open-source, token-efficient AI agent that provides high-capability agentic workflows and coding assistance with reduced costs via optimized caching and a minimal tool set.

  • Project
    snyk/agent-scan

    Snyk Agent Scan is a CLI security scanner that discovers AI‑assistant components (MCP servers, tools, and skill files) on a machine, sends them to Snyk’s analysis service, and reports scored risks such as prompt‑injection, secret leakage, and malicious code. It auto‑detects many agents (Claude, Cursor, Copilot, Gemini, etc.) across macOS, Linux and Windows, can run via `uvx` or a standalone binary, and offers both interactive and background (MDM) modes.

  • Project
    amplifthq/opentag

    A self-hosted framework that turns an AI coding agent into a persistent Slack teammate that executes tasks on a user-controlled local machine.

  • Project
    codeany-ai/open-agent-sdk-typescript

    An open-source TypeScript SDK for running autonomous AI agents in-process, featuring a vast library of built-in tools and support for multiple LLM providers.

  • Project
    kunchenguid/chrome-devtools-axi

    chrome‑devtools‑axi is an AXI‑compliant CLI that wraps the Chrome DevTools MCP server, giving LLM agents a low‑token, persistent, and ergonomically‑designed way to drive headless Chrome (navigate, click, fill forms, run JS, take screenshots, run Lighthouse, etc.). It keeps a bridge process alive across commands, uses TOON encoding to cut token usage ~40 %, and ships as an Agent Skill that agents can load on‑demand. Benchmarks show it uses the fewest input tokens, lowest cost, shortest duration, and fewest agent turns among seven browser‑automation setups, with 100 % task success.