Skip to content

Repository files navigation

elmah.io AI Plugin

Error monitoring and log management for your production apps, directly from your AI editor.

This plugin connects any MCP-compatible AI editor to the elmah.io MCP server, giving you and your AI agent direct access to production logs, error details, deployment history, and uptime monitors — without leaving your editor.

Supported Editors

Editor Manifest
Cursor .cursor-plugin/
Claude Code .claude-plugin/
Any MCP-compatible editor Point at https://mcp.elmah.io/mcp

Installation

Cursor

OAuth (recommended):

{
  "mcpServers": {
    "elmah.io": {
      "url": "https://mcp.elmah.io/mcp",
      "auth": {
        "CLIENT_ID": "ai-plugin"
      }
    }
  }
}

API key:

{
  "mcpServers": {
    "elmah.io": {
      "url": "https://mcp.elmah.io/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Claude Code

OAuth (recommended):

{
  "mcpServers": {
    "elmah.io": {
      "type": "http",
      "url": "https://mcp.elmah.io/mcp",
      "oauth": {
        "clientId": "ai-plugin"
      }
    }
  }
}

On first use, your editor will prompt you to sign in with your elmah.io account. Access tokens expire after 30 days; re-authenticate if you see auth errors.

API key:

{
  "mcpServers": {
    "elmah.io": {
      "type": "http",
      "url": "https://mcp.elmah.io/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Skills

Skill Description
debug-production-error Investigate a specific error by ID, message, or stack trace
post-deployment-check Verify production health after shipping code
investigate-frequent-errors Find and prioritize your highest-impact errors
check-uptime-status Check monitor status and investigate downtime events

Usage examples

Debug this production error: NullReferenceException in OrderService
Run a post-deployment check after my latest deploy
What are the most frequent errors in my production log this week?
Is my site up? Check uptime status

Agents

deployment-watchdog — A background agent that monitors for regressions for 30 minutes after a deployment. Alerts immediately on fatal errors or downtime; otherwise reports back with a final HEALTHY / UNHEALTHY verdict.

Start the deployment watchdog for version 2.4.1

MCP Tools Available

Once connected, the following elmah.io tools are available to your AI agent.

Read-only

  • organizations_list / organizations_get_details: list organizations and get quota/stats
  • logs_list / logs_get_details: list logs and get configuration details
  • messages_list_recent: fetch recent log messages with filtering
  • messages_list_frequent: get the most frequently occurring errors
  • messages_get: fetch a single message with full details (stack trace, custom data, etc.)
  • messages_count: count messages matching a query
  • deployments_list: list deployments with timestamps and versions
  • uptime_list / uptime_get_details: list uptime monitors and get incident history
  • heartbeats_list / heartbeats_get_details: list heartbeat monitors and check health
  • recommendations_list / recommendations_get_details: list and read automated suggestions for improving your monitoring setup, security and configuration
  • users_get_current: get the authenticated user
  • system_get_context: get system information like current date and time

Change data

These tools modify messages in your log. The server tells the agent to call them only when you have explicitly asked, and none of the plugin's skills or agents use them.

  • messages_fix / messages_open: mark a message as fixed, or reopen it
  • messages_hide / messages_unhide: hide a message, or unhide it
  • messages_bot / messages_unbot: mark messages as bot traffic (1-500 per call), or undo it

Requirements

  • An active elmah.io account
  • At least one log configured in elmah.io

License

Apache 2.0 — see LICENSE