convex-monitor

Watches for the next error or request event in a Convex app and reacts to it.

9.4k|1.5k|Updated Jan 3, 2026
One-click install
npx skills add https://github.com/openclaw/clawhub --skill convex-monitor
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: convex-monitor
Source: https://github.com/openclaw/clawhub/tree/main/.agents/skills/convex-monitor
Command: npx skills add https://github.com/openclaw/clawhub --skill convex-monitor

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Developers working on Convex apps waste time manually polling logs and dashboards to catch errors or feature requests. This Skill blocks on the next typed event—dev errors, production errors, or feature requests—so you can react immediately instead of constantly checking.

Core Features & Use Cases

  • Blocking Event Wait: Calls wait_for_event with project directory, event kinds, and timeout to race local error logs, deployment subscriptions, and Sentinel prod-error rows, returning the first to fire.
  • Typed Event Reactions: Decodes and fixes convex_error events, triages prod_error events via Sentinel, builds feature_request events, and loops on quiet heartbeats.
  • Poll-Loop Fallback: In harnesses without blocking MCP support (e.g., Copilot cloud), runs a poll loop with the same versioned event contract.
  • Use Case: While iterating on a Convex backend, let the agent block on the next dev error, automatically decode it, and apply a fix without you switching to the dashboard.

Quick Start

Ask the agent to watch my Convex project for the next error or feature request and react to it when one arrives.

Frequently Asked Questions about convex-monitor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I watch for the next error in a Convex app?

Call wait_for_event with your project directory, the event kinds you care about, and a timeout in milliseconds. It races local error logs, deployment subscriptions, and Sentinel prod-error rows, returning the first event to fire or a quiet heartbeat.

How to monitor Convex production errors automatically?

Subscribe to prod_error events, which come from Sentinel rows in your deployed cloud app. When one fires, triage it using the Sentinel workflow and then apply a fix. Production events require a deployed Convex cloud app with Sentinel configured.

Does Convex event monitoring work without a blocking MCP tool?

Yes. In harnesses without blocking MCP support, such as Copilot cloud, the skill runs a poll loop instead. The poll loop uses the same versioned event contract, so behavior and event shapes stay identical.

What event types can the Convex monitor react to?

It handles four typed events: convex_error or next_error for local dev errors, prod_error for production issues via Sentinel, feature_request for new requests to build, and quiet heartbeats when nothing fires within the timeout.

Why does the monitor return quiet events repeatedly?

A quiet event means no error or request fired before the timeout elapsed. The expected behavior is to loop and call wait_for_event again, keeping the watch alive until a real event arrives.