← Back to Discover

squirrelscan

Open-source command-line website auditor for coding agents: checks SEO, performance, security, accessibility and agent readiness, and returns fixes an agent can apply.

Testing & QACommand-line toolsMCP servers
Visit website ↗

What it does

squirrelscan crawls a website and reports problems across SEO, performance, security, accessibility and what it calls agent experience, such as llms.txt and AI crawler handling. Every finding comes with a concrete fix, written so a coding agent can act on it directly. It is designed around a loop with agents like Claude Code, Cursor and Codex. The agent runs an audit, applies fixes and re-audits until the score reaches 100. It can be driven from the command line, as an agent skill, through a local MCP server started with squirrel mcp, or through an API, and its CI-friendly exit codes let a build fail on a score or severity threshold. Local audits are free and need no account. Signing in adds cloud features paid for with monthly credits: headless browser rendering for JavaScript-heavy sites, AI page analysis, technology detection, dead-link checking, scheduled audits and shareable published reports. Reports come in seven formats, including a self-contained HTML file.

Capabilities

  • More than 300 audit rules across 22 categories
  • Fix guidance formatted for coding agents to apply
  • Runs as a CLI, an agent skill, an MCP server, a GitHub Action or a cloud service
  • Health score and letter grade in a shareable HTML report
  • CI exit codes to fail a build on score or severity
  • Cloud add-ons for browser rendering, AI page analysis and scheduled audits

When to use it

  • Letting a coding agent fix SEO, accessibility and performance issues in a loop
  • Gating pull requests or deployments on a site-quality score
  • Auditing a site locally without creating an account
  • Auditing JavaScript-heavy sites with cloud browser rendering

Alternatives to explore

  • Screaming Frog