Broken Mistral Crawler

crawler.yaml

# Before starting this, you need to create the mistral-environment.
#     $ cat ~/env.json
#     {"name": "crawler", "variables": {"visited": []}}
#     mistral environment-create env.json

---
version: '2.0'
name: crawler
description: Web Scraping Workbook

workflows:

  get_web_page:

    description: >
      Fetch a given URL. Store it. Extract all the links and queue them.

    input:
      - url

    tasks:

      limit_domain:
        on-success:
          - get_mistral_env: <% 'dougalmatthews.com' in $.url %>

      get_mistral_env:
        action: mistral.environments_get name='crawler'
        on-success:
          - mark_visited: <% not $.url in task(get_mistral_env).result.variables.visited %>
          - download_url: <% not $.url in task(get_mistral_env).result.variables.visited %>

      mark_visited:
        action: mistral.environments_update name='crawler' variables=<% {visited => [$.url] + task(get_mistral_env).result.variables.visited} %>

      download_url:
        action: std.http url=<% $.url %> allow_redirects=True
        on-success: extract_links

      extract_links:
        publish:
          urls: <% regex('((?:https?://|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:\'".,<>?\xab\xbb\u201c\u201d\u2018\u2019]))').searchAll(task(download_url).result.content) %>
        on-success:
          - echo_links
          - follow_links

      echo_links:
        action: std.echo output=<% task(extract_links).published.urls %>

      follow_links:
        with-items: suburl in <% task(extract_links).published.urls %>
        workflow: get_web_page url=<% $.suburl %>

maybe use env() to do this?