🥜 cli website audit tool (seo, security, perf, etc.) built for coding agents - https://xitter.cf/t.co/9CbqAa5ftJ
tree
Joined January 2026
- Tweets159
- Following29
- Followers51
- Likes56
Pinned Tweet
we also have one of those fancy videos .. intro to squirrelscan, a website audit tool for seo, performance, security and more that integrates into your coding agent 🚀🥜 🐿️
🥜💅
Your coding agent can write the fix. It still needs to find the website problem.
squirrelscan is an open-source CLI website QA tool for builders using coding agents.
It helps you audit a site and give your agent specific fixes by crawling it against 300+ rules across SEO, performance, security, accessibility, and agent experience.
Key features:
• Broad audit coverage – 300+ rules across 21 categories cover website QA in one CLI
• Agent experience checks – tests whether agents can discover, read, and act on your site
• MCP and skills support – connect the tool to coding agents and run audit-and-fix workflows
• Incremental crawling – uses ETag, Last-Modified, and content hashing, with checkpoints for resuming
• Flexible reports – export findings to console, JSON, HTML, Markdown, text, LLM, or XML formats
It’s open-source (MIT license).
Link in the reply 👇
Turns out detecting broken links is haaaard.
We shipped a link checker that cried wolf so often we had to rebuild it. Meet squirrel perch.
Thread on why it's hard now, and what we did about it 🧵
Pro users in cloud get always-on link monitoring: if a link goes dead, you get an alert. Email for now, MCP message alerts soon.
Coming soon: when a dead link has a copy in the web archive, we'll suggest the archived link so you can swap it in :)
The full write-up, with the honest caveats and the numbers:
squirrelscan.com/blog/squirr…
Heads up: the cli without signing in checks from your own machine and lots of sites block that, so expect false positives. Sign in and checks go through our cluster.
every squirrelscan audit now builds an entity map: your site's json-ld, collapsed into one graph, so you can see whether it declares one organization or a fresh one on every page.
and your agent can fix what it finds.
also new: noindex pages skip title/description/h1 checks (unless the homepage is noindex), plus false-positive fixes, and much more. big audits use about half the memory. cloud audits get include/exclude patterns and a sitemap-only scope.
curl -fsSL install.squirrelscan.com | bash
or squirrel self update
the launch post, with the whole loop: squirrelscan.com/blog/entity…