How the agent fetches Zendesk tickets, classifies them, searches Notion KB, and routes to Slack in five steps

A real working agent you can deploy

This repo uses a single SCALEKIT_USER_ID env var to simulate one user. In production, pass each user's real ID as the identifier on every Scalekit call, and send them an authorization link whenever their connector status is not ACTIVE.

Authorize, then orchestrate

connectors.py

Each connector wrapper checks its connected account at startup via get_or_create_connected_account. If Zendesk, Slack, or Notion is not ACTIVE, the agent surfaces a re-authorization link instead of failing silently. Scalekit refreshes tokens across all three connectors.

connectors.py

Copy

class Connector:
    def __init__(self, actions, connector_name, identifier):
        self.actions = actions
        self.connector_name = connector_name
        self.identifier = identifier

    def check_auth(self) -> bool:
        resp = self.actions.get_or_create_connected_account(
            connection_name=self.connector_name,
            identifier=self.identifier,
        )
        if resp.connected_account.status != "ACTIVE":
            link = self.actions.get_authorization_link(
                connection_name=self.connector_name,
                identifier=self.identifier,
            ).link
            logger.warning(f"Authorize here: {link}")
            return False
        return True

Fetch new Zendesk tickets

connectors.py

Classify via LLM or rule-based fallback

classifier.py

Search the Notion knowledge base

connectors.py

Route to Slack and update Zendesk

triage_engine.py

Agents that triage Zendesk tickets, without the auth plumbing

Three things you'd otherwise build: Zendesk OAuth for ticket access, Notion token storage for KB search, Slack identity for routing. All must be live when the ticket lands. Handled.

Install

Copy Copied

npx @scalekit-inc/cli setup
# select your coding agent
# enable scalekit MCP server on coding agent

terminal

Copy Copied

npx @scalekit-inc/cli setup
# select your coding agent
# enable scalekit MCP server on coding agent

terminal

Copy Copied

npx @scalekit-inc/cli setup
# select your coding agent
# enable scalekit MCP server on coding agent

terminal

Copy Copied

npx @scalekit-inc/cli setup
# select your coding agent
# enable scalekit MCP server on coding agent

terminal

Copy Copied

npx @scalekit-inc/cli setup
# select your coding agent
# enable scalekit MCP server on coding agent

Prompt

Copy Copied

You have the Scalekit authstack plugin installed (AgentKit + SaaSKit skills and MCP auth). Before coding, auto-load the AgentKit skills relevant to connectors (Zendesk, Slack, Notion) and use the Scalekit MCP server to validate tool calls instead of hand-written API logic.

Poll Zendesk for new and open tickets. Classify each by category and severity with an LLM (fall back to keyword rules if the API is unavailable). Search the Notion knowledge base for matching articles when the category calls for it. Post a formatted triage message to the category's Slack channel, then update the Zendesk ticket with tags, priority, and an internal note. Track processed tickets in a file-backed state store so restarts never re-triage.

Read-only steps (Zendesk fetch, Notion search) run without asking; ask before using write tools (Slack post, Zendesk update). Exit 0 on success, 1 on error, 2 if no new tickets.

Why choose Scalekit