The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes subprocess.run with shell=True in scripts/agent.py to execute commands defined within external YAML files. While this is intended to facilitate threat emulation tests, it represents a significant capability that executes arbitrary shell strings.
[DYNAMIC_EXECUTION]: The functions execute_atomic_manual and run_cleanup in scripts/agent.py dynamically construct shell commands at runtime. They perform string replacement on templates loaded from external Atomic Red Team YAML definitions before passing them to the system shell.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to download test definitions and execution frameworks from Red Canary's official GitHub repositories, specifically for the Atomic Red Team and atomic-operator projects.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill parses external YAML files via the load_atomic_tests and list_available_techniques functions in scripts/agent.py.
Boundary markers: None identified; the skill processes command strings directly from the atomic_tests data structure.
Capability inventory: The agent possesses the capability to execute shell commands and cleanup routines using subprocess.run.
Sanitization: While the skill uses shlex.quote to sanitize individual input arguments, the primary command skeleton is executed as-is from the parsed YAML content.

performing-threat-emulation-with-atomic-red-team