security(agent): Replace unsafe pyyaml loader with SafeLoader

[matheudev]

Background

Changes 🏗️

The default loaders in PyYAML are not safe to use with untrusted data. They potentially make your application vulnerable to arbitrary code execution attacks. If you open a YAML file from an untrusted source, and the file is loaded with the default loader, an attacker could execute arbitrary code on your machine.

This codemod hardens all yaml.load() calls against such attacks by replacing the default loader with yaml.SafeLoader. This is the recommended loader for loading untrusted data. For most use cases it functions as a drop-in replacement for the default loader.

Calling yaml.load() without an explicit loader argument is equivalent to calling it with Loader=yaml.Loader, which is unsafe. This usage [has been deprecated](https://github.com/yaml/pyyaml/wiki/PyYAML-yaml.load(input\)-Deprecation) since PyYAML 5.1. This codemod will add an explicit SafeLoader argument to all yaml.load() calls that don't use an explicit loader.

More reading

• https://owasp.org/www-community/vulnerabilities/Deserialization_of_untrusted_data
• https://github.com/yaml/pyyaml/wiki/PyYAML-yaml.load(input)-Deprecation

PR Quality Scorecard ✨

The changes from this codemod look like the following:

  import yaml
  data = b'!!python/object/apply:subprocess.Popen \\n- ls'
- deserialized_data = yaml.load(data, yaml.Loader)
+ deserialized_data = yaml.load(data, Loader=yaml.SafeLoader)

• Have you used the PR description template?   +2 pts
• Is your pull request atomic, focusing on a single change?   +5 pts
• Have you linked the GitHub issue(s) that this PR addresses?   +5 pts
• Have you documented your changes clearly and comprehensively?   +5 pts
• Have you changed or added a feature?   -4 pts
  • [] Have you added/updated corresponding documentation?   +4 pts
  • Have you added/updated corresponding integration tests?   +5 pts
• Have you changed the behavior of AutoGPT?   -5 pts
  • Have you also run agbenchmark to verify that these changes do not regress performance?   +10 pts

[netlify]

✅ Deploy Preview for auto-gpt-docs canceled.

Name	Link
🔨 Latest commit	7c28b3b
🔍 Latest deploy log	https://app.netlify.com/sites/auto-gpt-docs/deploys/65fd8963af24fc0008a7dc9b

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 45.51%. Comparing base (30bc761) to head (7c28b3b).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7035   +/-   ##
=======================================
  Coverage   45.51%   45.51%           
=======================================
  Files         139      139           
  Lines        6530     6530           
  Branches      917      917           
=======================================
  Hits         2972     2972           
  Misses       3408     3408           
  Partials      150      150           

Flag	Coverage Δ
Linux	45.40% <100.00%> (ø)
Windows	43.64% <100.00%> (ø)
autogpt-agent	45.48% <100.00%> (ø)
macOS	44.85% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Pwuts]

Good idea, I don't think we use the extra/dangerous YAML features anyways.