Skip to main content
skill CX Codex

convex-sentinel — Claude Code skill: what it does, how to install

Configures Sentinel error capture inside your own Convex deployment. Useful for tracking server failures and client React crashes without sending logs to third-party services.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add convex/agent-skills --skill convex-sentinel -a claude-code
Open on GitHub

What convex-sentinel does

This skill configures @convex-dev/sentinel to store production runtime issues directly in your Convex database. It captures server function failures, client React crashes, window.onerror events, and concurrency conflict signals. All data stays strictly inside your deployment.

The skill wires app.use(sentinel) into convex/convex.config.ts and configures a React error boundary with breadcrumbs. Secret redaction runs at write time by default through a deny list on secret names and value patterns. This keeps credentials out of the error tables before the agent reads them for model processing.

You can inspect logs through Convex CLI queries or listen to the prod_error event. It also supports an optional self-healing cron: the triage step tags recurring non-transient failures and passes them to ai-runner to open a pull request. Real error capture requires a live cloud deployment.

What it helps with

  • capture Convex server failures and client React crashes into internal tables
  • redact credentials and secret tokens at write time before storage
  • monitor runtime incidents via Convex CLI and the prod_error event
  • automate fix pull requests for recurring bugs with triage and ai-runner

Example prompts for the agent

  • Set up convex-sentinel in convex/convex.config.ts and wire the React error boundary
  • Inspect recent production errors recorded by Sentinel using Convex CLI
  • Enable the triage self-healing cron job to open fix pull requests with ai-runner

How to install convex-sentinel

1Skills CLI (Vercel)

npx skills add convex/agent-skills --skill convex-sentinel -a claude-code

2Manually: put the files in place

~/.claude/skills/convex-sentinel/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/convex-sentinel/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about convex-sentinel

How to install convex-sentinel in Claude Code?

Run npx skills add convex/agent-skills --skill convex-sentinel -a claude-code in your terminal.

Where is error data stored?

All logs and crash reports are written to a table inside your own Convex deployment. The data never leaves your infrastructure or goes to third parties.

How does convex-sentinel protect sensitive credentials?

Redaction runs at write time by default with a default-deny policy on secret key names and patterns. This prevents raw secrets from reaching LLM providers when the agent reads error data.

Source text

SKILL.md · Apache-2.0
---
name: convex-sentinel
description: "Set up Sentinel production error capture in your own Convex deployment."
---

<!-- GENERATED from convex-agents content/capabilities/sentinel.json — do not edit by hand. -->

# Capture production errors in your own deployment

Install `@convex-dev/sentinel` to capture production errors (server function failures, client JS/React crashes, OCC and scale signals) into a table in the user's OWN deployment, redacted at write time, then react to new ones. Data never leaves the user's deployment.

## Workflow

1. Install the component: `app.use(sentinel)` in `convex/convex.config.ts`.
2. Wire the client SDK: a React error boundary plus `window.onerror`/`unhandledrejection` and breadcrumbs.
3. Redaction runs at write time and is on by default (default-deny on secret key names and value patterns).
4. Read recent errors with the Convex CLI (`convex data`, `run-once-query`); react to new ones via the monitor's `prod_error` event.
5. Optionally enable the self-healing cron: `triage` classifies each error and, for recurring non-transient ones, hands it to ai-runner to open a fix PR.

## Rules

- Redaction is mandatory and on by default — never store raw secrets; the agent's reads reach the model provider.
- Data stays in the user's deployment; never send it to a third party.
- Sample and cap to control volume and cost.
- Capturing PROD errors needs a deployed cloud app; install works anonymously.

© convex · Apache-2.0 · convex/agent-skills

Similar skills

Everything for Backend and APIs →

From the same set32

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor