KEEPitALIVE

Production systems fail, and generic LLMs often hallucinate status codes or suggest the wrong monitoring type for private infrastructure. This skill provides a structured framework for configuring and managing KEEPitALIVE monitoring through AI agents. It ensures your agent selects the correct monitoring strategy—whether that is active polling for public APIs or heartbeats for internal cron jobs—while adhering to security best practices and official API contracts.

What it does

• Configure monitors by selecting between HTTP, API, Browser, Ping, TCP, DNS, and Game Server types based on service requirements.
• Instrument heartbeats for private-network jobs, backups, and cron tasks to detect failures or missed runs without exposing internal ports.
• Automate incident response by retrieving check history, observing status transitions, and drafting evidence-based post-mortems.
• Manage triggers to configure webhooks and notifications based on state changes or specific event payloads.
• Audit reliability by generating reports from actual observation windows, latency data, and recorded downtime.

How it works

1. Discovery: The skill retrieves the latest agent guide and OpenAPI specification to ensure endpoint names and payloads are current.
2. Classification: It analyzes the target service to determine if it requires active polling or a passive heartbeat/event receiver.
3. Configuration: The agent plans the monitor setup, including intervals and alert destinations, requesting explicit approval before resource mutation.
4. Operation: It executes verified MCP tools or REST calls using scoped API keys, adhering to least-privilege principles.

Frameworks & tools

This skill interfaces with the KEEPitALIVE REST API and the @keepitalive/mcp server. It is designed for use within environments like Claude Code or Cursor that support MCP or outbound HTTP requests.

Why this beats prompting it yourself

Generic prompts often fail to distinguish between active polling and heartbeats, leading to "false green" status for private jobs. This skill enforces precise monitoring semantics, prevents the fabrication of uptime metrics, and ensures that sensitive heartbeat tokens are never committed to version control.

Use cases

• Monitor a public REST API including specific JSON response body assertions.
• Track success/failure of nightly database backups running in a private VPC.
• Notify the team via webhook when a GitHub Actions deployment fails.
• Generate a weekly reliability report based on verified monitoring data.

Known limitations

Requires an active KEEPitALIVE account and a scoped API key. Does not independently host checks; it manages the KEEPitALIVE service. Destructive operations and billing changes require manual confirmation.