The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of external applications via the Python subprocess module.
Evidence: The execute_safely function in modules/execution-patterns.md uses subprocess.run to launch command strings constructed from service configurations.
[PROMPT_INJECTION]: The skill's command construction logic is vulnerable to indirect prompt injection by embedding untrusted user data into command templates.
Ingestion points: The prompt and files arguments are consumed by the build_command function in modules/execution-patterns.md.
Boundary markers: Not utilized; input is directly interpolated into strings such as "{command} -p {prompt}".
Capability inventory: High-risk capability identified: arbitrary command execution on the host system.
Sanitization: The skill uses shlex.split which mitigates shell-level character escaping issues but does not sanitize the content to prevent it from being interpreted as malicious flags or parameters by the target CLI tools.

service-registry