the cli

Run the scan from your terminal.

Run the agent-readiness scan in your shell and your CI. crwl reads a site the way an agent does and prints a score, the gaps, and what to fix. No account, no key.

$curl -fsSL crwl.sh/install | sh

or: go install github.com/crwl-sh/crwl-cli/cmd/crwl@latest

scan a site

Point it at any host. crwl crawls what it can discover, then prints the report.

crwl scan northwind.dev          # scan and print the report
crwl scan northwind.dev --pages 1   # homepage only, fastest
crwl scan northwind.dev --json      # machine-readable output

gate a build

Set a floor with --min. crwl exits non-zero below it, so a readiness regression fails the job like any other check.

# .github/workflows/agent-readiness.yml
name: agent readiness
on: [push]
jobs:
  crwl:
    runs-on: ubuntu-latest
    steps:
      - run: curl -fsSL crwl.sh/install | sh
      - run: crwl scan ${{ github.event.repository.homepage }} --min 80

flags

--min <0-100>: Minimum score to exit 0. Below it, crwl exits 2 so a CI job fails.
--pages <n>: Max pages to crawl. 0 crawls everything discovered; 1 is the homepage only.
--json: Print the full report as JSON instead of the terminal view.
-o <file>: Also save the report to a file (JSON if the name ends in .json, else text).
--timeout <dur>: Per-request timeout, e.g. 10s. Default 10s.
--no-color: Disable ANSI colour in the terminal report.

exit codes

0: Scan ran and the score met --min (or no threshold was set).
2: Scan ran but the score was below --min. Use this to gate a build.
1: The scan could not complete (network, DNS, or an unreadable target).

open source

The crwl CLI is open source: Apache-2.0, no third-party dependencies. Run the same readiness checks locally and in CI. Read it, fork it, or open an issue.

View on GitHub ↗Or scan from the web →