I soft-launched a project I've been working on and using for months now in my coding agent workflows - squirrelscan. It is a cli auditing tool for website that checks seo, security, performance and 17 other categories over 150+ rules.
Previously I was running scans manually and copying results back into claude code and making fixes - which was a bit slow and laborious. The big unlock was the launch of agent skills and the broad support and adoption it has seen. Now I can run a skill from within my website project, have squirrel run an audit and report all the issues back to the llm to get fixed.
This loop works remarkably well - and I've been refining both the tool and the skill definition to ensure complete coverage and fixes across all categories. Watching it work kinda feels like magic :)
In 5 days it's already at over 5k installs via vercel's skills marketplace at skills.sh where it's ranking in the top 10 overall for installed skills.
There are now 150+ rules in 20 categories - all the usual stuff like robots/sitemap validation, title and desc length, parsing and validating schemas (and alerting when they're not present but should be), performance issues, security, E-E-A-T characteristics, a11y etc. but some of the more unique ones that you probably haven't seen are:
leaked secrets - a common vibe-code gotcha, it detects over 100 leaked secret types
video schema validation - I watched claude auto-create and include a thumbnail and generate a11y captions based on this rule being triggered
NAP consistency - it'll detect typos and inconsistencies across the site
Picks up render blocking and complicated DOM trees in performance rules
noopener on external links (find this all the time)
warns on public forms that don't have a CAPTCHA that probably should to prevent spam
adblock and blocklist detection - this is currently in the beta channel. it detects if an element or included script will be blocked by adblock, privacy lists or security filters. this came up because we had a webapp where elements were not displaying only to find out after hours of debugging that it was a WAF blocking a script.
squirrel is based on a crawler I've written myself and used for years so it's robust - with smart discovery and caching. I've had many examples of where other skills or tools would pass a website for SEO / audit only for squirrel to give it an F and point out 30+ errors. The coverage matches and exceeds some of the more well known and paid enterprise auditing tools - but there is a lot more I want to do!
To get started visit the website and follow the steps to install
You can run it on the command line to get markdown / html reports for humans, or open up your preferred coding agent and trigger it with /audit-website
There are many more features planned! Multiple releases and updates every day at the moment and you can follow updates on the recently launched twitter profile - thanks!
Congrats on the launch, curious which metrics are giving you the most surprises?