Smoke Test Until Clean
Walk through your app's critical flows in a real browser, capture console errors and failed requests, fix them, and repeat until every flow runs clean.
Copy the Smoke Test Until Clean loop
Stops when every critical flow completes with no console errors and no failed network requests, or after 8 iterations. Edit the settings block, then paste the whole prompt into your agent.
You are running the "Smoke Test Until Clean" loop. Settings (edit before running): - Base URL: http://localhost:3000 (local or staging only) - Test account: test-user@example.com / use the seeded test password - Flows: 1. Sign up with a new random test email and land on the dashboard. 2. Create a project, rename it, and delete it. 3. Open settings, change the display name, reload, and confirm it persisted. - Max iterations: 8 Goal: every flow completes with zero console errors and zero failed network requests (4xx or 5xx, except expected ones you list). Each iteration: 1. Run every flow in a fresh browser context. At each step, record console errors and warnings, failed requests (method, URL path, status), and a screenshot if the step fails. 2. List problems in order of appearance, grouped by flow. 3. Take the first problem. Find the code responsible using the stack trace, request path, and server logs. 4. Fix it in the source. Restart the dev server if needed. 5. Re-run all flows. Log: iteration, problems before and after. Stop when: - All flows pass with no console errors or failed requests. Success. - You reach the max iteration count. Rules: - Only use the base URL above and the test account. Never use real user accounts, real payment details, or production. - Do not suppress errors by catching and ignoring them, or by removing console logging. - Do not change the flows to avoid a broken step. - Clean up any test data you created if the app offers a way to do so. Final report: flow results, each problem with its cause and fix, and screenshots of anything still failing.
When this loop fits
- Before a release, to click through sign-up, checkout, and settings the way a user would.
- After a framework upgrade, when pages render but something in the browser console is red.
- When unit tests pass but you have no end-to-end suite yet.
What happens on each pass
The agent repeats these steps until the stop condition holds or it runs out of iterations.
- Define the flows. Write each critical flow as a short list of user actions and the expected result.
- Run each flow. Drive the browser through the steps while recording console messages, failed requests, and screenshots at each step.
- Diagnose. Match each error to the code that caused it, using stack traces and request URLs.
- Fix and re-run. Fix one cause, then re-run every flow so fixes do not break other paths.
MCP servers for this loop
The 6 strongest of 706 matching servers, ranked by how closely they fit and then by trust score.
- 72Trust score 72 of 100, Good
moontide
by konni-arcadia · JavaScript
Early: Claude steers live music by writing Strudel code played in a browser page
RegistryLicenseQualityMaintenanceLicense: C. AGPL-3.0 is a strong copyleft license: derived works must use the same license.Quality: Not graded. No README or tool list has been captured yet.Maintenance: A. Last commit or release 21 days ago. - 70Trust score 70 of 100, Good
Fullbleed PDF Engine
by fullbleed-engine · Rust
Create, preview, inspect, verify, and compile deterministic PDF documents with Fullbleed.
RegistryLicenseQualityMaintenanceLicense: A. MIT is a permissive open source license.Quality: A. 91/100 for listing completeness. Lost points: short summary.Maintenance: A. Last commit or release today. - 70Trust score 70 of 100, Good
mcp
by webdriverio · TypeScript
MCP server with WebdriverIO for browser and mobile app automation (iOS/Android via Appium)
RegistryLicenseQualityMaintenanceLicense: A. MIT is a permissive open source license.Quality: Not graded. No README or tool list has been captured yet.Maintenance: A. Last commit or release 11 days ago. - 70Trust score 70 of 100, Good
Kitsune
by junyeong-nero · Python
Let your coding agent use the Firefox, Zen, or LibreWolf browser you already have open.
RegistryLicenseQualityMaintenanceLicense: A. MIT is a permissive open source license.Quality: Not graded. No README or tool list has been captured yet.Maintenance: A. Last commit or release 4 days ago. - 69Trust score 69 of 100, Good
browser-gateway
by browser-gateway · TypeScript
AI browser tools with Chrome CDP. Navigate, screenshot, interact. Multi-provider failover.
RegistryLicenseQualityMaintenanceLicense: A. MIT is a permissive open source license.Quality: Not graded. No README or tool list has been captured yet.Maintenance: A. Last commit or release 22 days ago. - 69Trust score 69 of 100, Good
Proxy MCP for Playwright, Figma, DevTools, Shopify Dev, Context7 — fewer tokens, caching
RegistryLicenseQualityMaintenanceLicense: A. MIT is a permissive open source license.Quality: Not graded. No README or tool list has been captured yet.Maintenance: A. Last commit or release 8 days ago.
How to run this loop
- Connect the tools. Install a server the loop can use, for example moontide, or rely on your agent's built-in terminal and file access.
- Edit the settings block. Replace the example commands, targets, and limits with your project's own. Keep the iteration cap low on the first run.
- Paste it into an agent session. Start a fresh session in Claude Code, Cursor, or VS Code agent mode, in the repository root, and paste the prompt.
- Review before you merge. Read the final report and the diff. The loop leaves its work for you to review rather than shipping it.
Smoke Test Until Clean questions
When does the Smoke Test Until Clean loop stop?
It stops as soon as every critical flow completes with no console errors and no failed network requests. It also stops after 8 iterations by default, or earlier when the prompt's own stop rules say it is stuck, and then reports what changed and what is left. Change the cap in the settings block at the top of the prompt.
How is this different from an end-to-end test suite?
An end-to-end suite is code you maintain. This loop is a prompt: the agent drives the browser from plain-language flows, so you get coverage on day one. Once the flows are stable, ask the agent to turn them into Playwright tests so they run in CI.
Can I point this loop at production?
Do not. The loop signs up users, creates and deletes records, and repeats that several times. Use a local environment or a staging deployment with a seeded test account, so nothing it does touches real customers or data.