Broken TypeScript sample in your dynamic monitor creation docs

[David19876543210]

I ran a checker over your docs and 1 of the TypeScript samples doesn't compile
against checkly@8.23.1. On the Dynamic monitor creation > Similar checks from
a list of targets page:

new Request({ ..., skipSsl: true })

// TS error: Object literal may only specify known properties, but
// 'skipSsl' does not exist in type 'Request'. Did you mean to write
// 'skipSSL'?

Casing mismatch — the real field is skipSSL.

I think it's drift between the published package and the docs rather than a
mistake in the sample.

I'm building a small CLI that catches this in CI. Two questions if you have a
minute: does this kind of thing get reported to you by users, and is it
something anyone currently owns?

Happy to share the full JSON report if that's useful — just ask.

David

[sorccu]

Thanks for the report @David19876543210! We'll take care of that.

Yes, our users sometimes do report such issues, but we mostly catch them during internal use. And now with AI, agents usually catch outdated or incorrect documentation very accurately when new features or fixes are being built. We don't have anyone who directly owns all the documentation.

[David19876543210]

Appreciate the honest answer about docs ownership. Ran the same check across the rest of your docs and found 3 more, on the create-multiple-monitors and uptime-monitoring guides (plus a second copy of the original skipSsl typo on the same page). Happy to send the full report over, free either way. Who should it go to?

[sorccu]

You were right, there were a whole bunch of issues in there. Embarrassing to say the least.

I used Claude to look for any issues and verified the fixes personally just to make sure. It seemed to do a fairly thorough job. I'm not sure if this is interesting to you, but it might be a useful data point: did Claude manage to find all the issues your tooling found?

[David19876543210]

Good question, and happy to give a precise answer rather than a vibe.

I rechecked both states of docs#499 directly — the base commit and the merge commit — with snippetcheck against checkly@9.1.0, scanning the actual repo rather than the live docs site. Against the same source:

• Your Claude pass fixed all 34 real compile errors that existed at the base commit, across 4 pages. My tool finds the identical 34 — not more, not fewer.
• After the fix: zero errors remain, and nothing new got introduced.
• Claude also caught one thing my tool structurally can't see on its own: creating-your-first-test.mdx had retryStrategy: { type: 'linear' } (should be 'LINEAR') hiding behind an import from the old @checkly/cli package name instead of checkly. My tool matches import strings exactly, so it silently skips anything still importing under the pre-rename name — that's a real gap in my tool, not a strength of mine over yours.

One correction on my end, in the interest of accuracy: my original report and follow-up undercounted badly — not because your Claude pass out-checked me, but because 30 of those 34 errors live on detect/uptime-monitoring/cli-configuration.mdx, a page that's entirely absent from your own llms-full.txt export (checked both before and after the fix — it's just not in there). I was scanning that export, not your repo, so I never saw most of it. Scanning the actual .mdx source finds it immediately.

So: on equal footing, your pass and my tool agree exactly, and your pass beat me on one edge case. Thanks for the real comparison — genuinely useful data point.

[sorccu]

Sounds good, thanks for bringing this up!