agent-slackbot_skill

Overview

This skill lets AI agents and humans interact with Slack workspaces using standard bot tokens (xoxb-). It provides a TypeScript CLI to send and manage messages, read channels, and handle reactions — optimized for server-side and CI/CD workflows. It supports multiple bots across workspaces and outputs machine-friendly JSON by default.

How this skill works

The CLI authenticates using Slack Bot tokens you obtain from the Slack App configuration and validates tokens before storing them. Commands cover messaging, threaded replies, listing and inspecting channels and users, and adding/removing reactions. Multiple bot profiles can be stored and switched, and CI pipelines can use environment variables instead of stored credentials. All commands return structured JSON (with a --pretty option for human-readable output).

When to use it

• Automate notifications or alerts from CI/CD pipelines.
• Let an AI agent post updates, summaries, or responses in Slack channels.
• Manage bot-driven moderation, reactions, or simple workflows across multiple workspaces.
• Run scripted channel monitoring or periodic workspace summaries.
• Operate without a desktop user token — server-side integrations and headless environments.

Best practices

• Use dedicated bot tokens (xoxb-) created in Slack App settings and grant only required scopes (chat:write, channels:read/history, reactions:write/read, etc.).
• Store tokens securely; prefer environment variables in CI/CD and keep local credentials file permissions restricted to owner only.
• Use the --bot flag for one-off commands to avoid switching global active bot when scripting.
• Invite the bot to private channels before attempting reads or posts; bots can only edit/delete their own messages.
• Handle rate limits by detecting slack_webapi_rate_limited_error and applying exponential backoff; the tool includes retries.

Example use cases

• Send deployment or alert notifications from CI to a specific channel using a dedicated deploy bot.
• Monitor a channel for new messages and trigger automated summary posts or external webhooks.
• Program an agent to reply in thread to user queries or to add emoji reactions as acknowledgement.
• Run a scheduled workspace summary that lists channels, active users, and recent bot posts.
• Use multiple bots for segmented responsibilities (alerts, deploys, reports) and switch via the CLI or per-command flag.