The Agent Skills Directory

[COMMAND_EXECUTION]: The file deploy.py executes shell commands using subprocess.run(script, shell=True). This is the core functionality described in the skill's purpose.
[PROMPT_INJECTION]: The skill is subject to an indirect prompt injection surface because it builds a shell command by interpolating the service and env arguments directly without sanitization.
Ingestion points: The service and env string arguments in deploy.py.
Boundary markers: Not present; the strings are directly formatted into the command line.
Capability inventory: Arbitrary command execution via subprocess.run (found in deploy.py).
Sanitization: None; the script does not validate or escape the input strings before they are used in the shell.

deploy-service