How to use the Configure Notifications agent skill

What is an agent skill and what does this one do?

An agent skill is a folder of instructions, held in a SKILL.md file, that an AI agent reads when a request matches the skill. Instead of you writing the setup steps yourself, the agent follows the skill's steps and asks you questions along the way.

This skill configures notification integrations. Once it is set up, OMC can message you when a session ends, when Claude is waiting for your response, or when background tasks complete. During setup the agent asks for your bot token and chat ID, saves the settings to a JSON configuration file in your Claude config folder, and can send a test message to confirm the connection works.

• Telegram: the most detailed path in the skill file, with step-by-step bot creation, chat ID lookup, message format choice, event selection, config writing and a test message.
• Discord: listed as an option in the routing section (webhook or bot token plus channel ID). The file does not document its steps in detail.
• Slack: listed as an option in the routing section (incoming webhook URL). The file does not document its steps in detail.

How do you install the configure-notifications skill?

Download the ZIP for this skill from this page and unzip it. The skill is a single file, SKILL.md. Put the unzipped folder in one of two places, depending on whether you want it for every project or one project only.

For all your projects, place it in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.

If you use the Claude apps instead of the command line, you can upload the skill under Settings → Capabilities → Skills.

• Personal (all projects): ~/.claude/skills/
• Single project: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills, then upload

How do you trigger the skill with a natural request?

You do not need a special command. Ask in plain language and the agent will match your request against the skill's trigger phrases. The skill's triggers include "configure notifications", "setup notifications", "configure telegram", "setup telegram", "telegram bot", "configure discord", "setup discord", "discord webhook", "configure slack", "setup slack" and "slack webhook".

The skill picks a provider from your wording. If your request names Telegram, Discord or Slack, it follows that section. If you name no provider, the agent asks you which service you want.

• "Set up Telegram notifications so I get a message when my session ends."
• "Configure Discord notifications for my long-running tasks."
• "Setup Slack notifications."
• "Configure notifications" (no provider named, so the agent asks which service to use)

What do you need before you start?

For Telegram, you need a bot token from @BotFather and a chat ID. The skill explains how to get both. To create a bot, open Telegram, search for @BotFather, send /newbot, then choose a name and a username. BotFather returns a token. To find your chat ID, send /start to your new bot and check the getUpdates response. Personal chat IDs are positive numbers and group chat IDs are negative numbers.

The skill also relies on two command-line tools: jq, which reads and writes the JSON config file, and curl, which calls the Telegram Bot API. For Discord or Slack you need webhook or bot credentials, but the skill file does not detail how to obtain them.

• jq installed on your machine
• curl installed on your machine
• A Telegram account and a bot token from @BotFather
• A Telegram chat ID, found by messaging your bot
• For Discord or Slack: webhook or bot credentials (not detailed in the file)

What happens during Telegram setup?

The agent first checks whether a configuration already exists in your config folder. If it finds Telegram settings, it shows them and asks whether you want to update or reconfigure. Otherwise it asks you questions one at a time: the bot token (checked for the form digits:alphanumeric), the chat ID (must be a number), the message format (Markdown or HTML), and which events should trigger alerts.

Events you can choose from are session end, input needed (Claude is waiting for your response), session start, and session continuing (when a persistent mode keeps the session alive). The default selection is session end and ask-user-question. Settings are written to ${CLAUDE_CONFIG_DIR:-~/.claude}/.omc-config.json, and any events you did not select are disabled.

• Bot token: must match digits, colon, then alphanumeric characters
• Chat ID: a number; positive for personal chats, negative for groups
• Format: Markdown (recommended) or HTML
• Events: session end, input needed, session start, session continuing

How do you test the setup and fix common errors?

After writing the config, the agent can send a test message to your Telegram chat and reports whether it succeeded. If the request fails, the skill lists the most common causes. A 401 Unauthorized response means the bot token is invalid. A 400 Bad Request with "chat not found" means the chat ID is wrong or you have not sent /start to the bot. A network error means you should check your connection to api.telegram.org.

You can also skip the wizard and set environment variables in your shell profile. The skill says the notification system detects OMC_TELEGRAM_BOT_TOKEN and OMC_TELEGRAM_CHAT_ID without needing the config file.

• 401 Unauthorized: bot token is invalid
• 400 Bad Request: chat not found: wrong chat ID, or /start not sent to the bot
• Network error: check connectivity to api.telegram.org

Safety notes: scripts, network and credentials

This skill runs shell scripts and makes network calls. It runs jq and curl commands on your machine and sends requests to the Telegram Bot API. The bot token and any webhook URL are credentials that give control of the bot or channel, so treat them like passwords.

The skill stores the Telegram bot token in a plain JSON config file, not an encrypted store. Do not paste a token or webhook URL into an AI chat if you are not comfortable with that. Review the SKILL.md file before you install it, and check the config file afterwards.

• Runs scripts: yes (bash and jq on your machine)
• Uses network: yes (calls to api.telegram.org)
• The Telegram token is saved in plain text in the config file
• Rotate the token with @BotFather if you think it has been exposed