Codebase Autopilot

React Doctor

A daily guardian that scans your React app, fixes the worst issue, and opens a PR.

Easy startShort cycleOpen loopVisible first run

Setup in Loopany

How a run flows

On schedule

Every day · 6am

Step 1 · Scan

Scan the app

Step 2 · Prioritise

Pick one issue

Isolated git worktree · off main

Step 3 · Fix

Fix in isolation

Step 4 · Verify

Verify the fix

Step 5 · Ship

Open a PR

Board

PR board

Metric

Health score

open loop — repeats on the next tick, picking up where it left off

Prerequisites

react-doctorrequired
The scanner the loop runs each morning — fetched via npx, nothing to install. See what react-doctor checks
Verifier setup for code changesoptional
A repo-specific harness so every fix is proven against the real app before a PR. Set it up with the verifier-setup skill

How this loop works — schedule, exit, notifications

Suggested schedule: Daily · ~6am
Loop type: Open loop — an ongoing monitor with no finish line.
Exit condition: Runs indefinitely (no finish line).
Notifications: On by default — you're alerted on each run's outcome (completion or failure). Silence with notify: never.
Ease of starting: Runs with almost no setup — confirm a detail or two and go.
When you see value: The first scan fixes the worst issue and reports a health score the same morning.

Over the last 3 months, React Doctor opened 35 fix PRs against our production app. Missing JSX keys, a locale-sensitive month bug, an impure state updater, unlabeled form controls. 31 are merged and 4 are still open. Most took me less than a minute to review. A few needed three to five.

The PR list, filtered to the loop's own work: 31 closed, 4 open, one defect class per PR

Keeping each PR to one kind of defect made this work. I can review one of those over coffee. Spread the same work across 40 files under "lint fixes" and it'll sit there, even when the patch is right. Small PRs move.

For now, I'm still clicking merge. If it keeps sending clean diffs, we'll give it more room. It can earn that.

The loop runs at 6am, so the PR's waiting when I wake up. At 2pm it would be an interruption. In the morning it feels like the codebase tidied itself up.

The prompt this runs

Set up a daily loop that keeps this React project healthy. Every morning around 6am my local time, scan the project with `npx react-doctor@latest`, pick the single most severe issue it reports, fix it, and verify the fix before proposing it. Do all the work in a fresh git worktree created off the main branch and located outside the loop folder - the loop folder is a synced content home, not a workspace - so neither my working checkout nor the sync is polluted. Open a pull request for the fix - but if a previous React Doctor PR is still open (unmerged), don't open another one that day; just refresh the existing PRs' statuses and still report the score. Track the pull requests as one markdown card per PR, each with front matter `type: open` or `type: merged` so a kanban board can follow them from open to merged, and report react-doctor's health score every day as a metric. Set up the loop's dashboard at creation: a kanban of the PR cards plus a chart of the health score over time.

This is the exact intent appended to the bootstrap when you create the loop — your coding agent reads it, then proposes cadence and config and confirms with you before creating anything.