The Agent Skills Directory

[SAFE]: The skill is primarily instructional, offering best practices for testing external system contracts before implementation.
[COMMAND_EXECUTION]: The document includes examples of standard development commands using uv, curl, docker, and unzip. These are used benignly for manual verification of response codes, library defaults, and build artifacts.
[EXTERNAL_DOWNLOADS]: The skill mentions fetching libraries via uv and using the official redis:7-alpine Docker image. These are standard practices in development workflows and do not involve untrusted or suspicious sources.
[DYNAMIC_EXECUTION]: The Python examples demonstrate manual monkeypatching of the requests library to inspect outbound traffic, which is a legitimate debugging technique for developers.

verifying-external-behavior