The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/generate_doubles.py contains a generate_script method that takes user-provided content (via the --content argument) and writes it directly into a bash script file.
Evidence: In scripts/generate_doubles.py, the generate_script method constructs a script with script_content = f"#!/bin/bash\n...\n{template}\n..." where template is the args.content value.
[REMOTE_CODE_EXECUTION]: The generate_script method in scripts/generate_doubles.py explicitly changes the file permissions to make the generated script executable (file_path.chmod(0o755)), creating a vector for immediate execution of arbitrary commands if an agent or user runs the resulting file.
Evidence: file_path.chmod(0o755) in scripts/generate_doubles.py following the write operation.
[DYNAMIC_EXECUTION]: The skill facilitates the creation of executable files from text strings at runtime, which is a high-risk pattern for supply chain or injection attacks if the input content is not sanitized.
Evidence: The --type script option in the CLI interface of scripts/generate_doubles.py maps directly to the creation of executable shell files.

generating-test-doubles