Developer tools
Catch it before release
Six tools that each make one pre-release check repeatable: a weak connection, a save that doesn't stick, a sitemap that shrank, a form that posts nowhere, public browser settings and a messy issue report.
Most release bugs are boring: a save that doesn't survive a reload, a contact form that posts nowhere, a sitemap that lost half its URLs in a refactor and went unnoticed for a month. None of them is hard to catch once you look. Remembering to look before every release is the hard part. Each tool here turns one of those checks into something you can run again, and most of them fit in CI.
Test under real conditions
Reproduce a weak connection, or a save that doesn't stick after a reload.
Bean Network Tester
Why it belongs here
Waiting for a bad hotspot to reproduce a timeout is no way to test. Bean Network Tester adds delay, packet loss, bandwidth caps or dropped connections to a chosen process, address or port on 64-bit Windows 10 and 11, so a flaky network becomes a condition you can set. One detail worth knowing: target by process name and the rules follow a restarted app, target by process ID and they don't. It needs administrator rights, and an unfiltered session slows the whole machine. The program wasn't run for its Note.
Read full NoteNote published Sep 2, 2026App Crash Lab
Why it belongs here
An app can say "Saved" and still bring back the old value on the next reload. App Crash Lab is built for that gap. Point it at a local save flow, tell it which values must survive, and it checks them and hands you a standalone Playwright test to keep. It's MIT licensed, needs Node 22 or newer, and makes no model calls or account requests. It only protects the values you name, so a vague expectation gives you a vague test. Its Note covers the documentation; the CLI wasn't run.
Read full NoteNote published Oct 11, 2026
Check what you're about to ship
A shrinking sitemap, a dead form and loose browser settings are all cheap to catch before release and annoying after.
Sitemap Cohort Auditor
Why it belongs here
Sitemap Cohort Auditor compares the URLs two sitemaps declare, so a release that quietly drops pages or switches host shows up before it ships. A small policy file can fail the build on host, count or removal rules, and the JSON output and exit codes slot into CI. It reads declarations without crawling, so it can't tell you whether a page loads or gets indexed. The CLI wasn't run for its Note.
Read full NoteNote published Sep 12, 2026Static Form Inspector
Why it belongs here
Static sites are full of forms that look finished and send nothing. Static Form Inspector reads pasted HTML and flags empty actions, placeholder endpoints, contact forms that use GET and fields with no name, and a GitHub Action puts the same checks in CI. It's honest about its limit: a clean result doesn't prove the form delivers, so still send one real submission. Its maker also sells an optional repair service.
Read full NoteNote published Aug 21, 2026DeployPass
Why it belongs here
A page can load perfectly and still ship with weak security headers or loose cookie flags. DeployPass runs 19 passive checks on a public URL, covering headers, cookies, cross-origin settings and exposed frontend files, and groups what it finds by severity. It's free during its beta with no signup. Logged-in routes and server logic are outside what it looks at, so a passing report covers those checks and doesn't prove the app is secure. No scan was run for its Note.
Read full NoteNote published Sep 2, 2026
Triage a bug report
Pin down what an issue actually shows before anyone starts on a fix.
Public Issue Evidence Capsule
Why it belongs here
Before a maintainer commits to fixing an issue, it helps to know what's actually been shown. Public Issue Evidence Capsule is a GitHub Action that records one public, non-security issue's metadata, labels and linked pull requests as deterministic Markdown with a content hash, keeping reported claims separate from anything reproduced. It doesn't run the project, reproduce the bug or suggest a fix, and it uses no AI at runtime. The Action wasn't run for its Note.
Read full NoteNote published Aug 26, 2026
How these projects were chosen
Every pick has a published Project Note and covers a different check, so the groups follow the job instead of sorting tools from best to worst. Most of these Notes draw on public documentation, and where a tool wasn't installed or run, the entry says so. Where public information ties several tools to one maker, this edition includes only one. Payment, traffic and clicks play no part in what's included or in the order.
For developers, QA testers and site maintainers who keep meeting the same kind of failure and want a narrow, repeatable check for it, and who will confirm the tool's limits in their own setup.
A review can leave the picks as they are. When a published selection or its guidance changes, a short note here explains why.