agent-slack

This Skill addresses the challenges AI Agents face when efficiently reading, writing, searching, and automating message management within Slack. In practical development, AI Agents frequently need to communicate with teams, retrieve context, and send notifications. However, traditional API calls are cumbersome and incur high Token costs, leading to expensive large language model operations and slow response times.

Usage is straightforward: simply install it globally via Bun or npm, then operate it directly through the command line or natural language instructions. For example, tell your AI "Check recent messages in #general," and it will automatically invoke agent-slack to retrieve results in a compact JSON format. Sending messages, searching files, managing channels, and orchestrating workflows can all be done with a single command. It also supports human-in-the-loop browser editors (message compose) and native Slack draft management.

It is ideal for teams or individuals who need deep Slack integration into their AI workflows. This is especially true for developers already using AI coding agents like Claude Code, Codex, or Cursor, as well as for DevOps engineers and project managers who need to batch-manage messages or automate project status synchronization.

It is recommended to use this tool in non-CI environments to fully leverage its human-in-the-loop interaction capabilities. Note that zero-config authentication relies on Slack Desktop; a manual token import is required for pure server environments. Additionally, it primarily focuses on text and file operations and does not support advanced features like video calls.