The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the omc ask command to interact with local AI providers (Claude, Codex, Gemini). This involves spawning a subprocess with user-supplied arguments.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection (Category 8) by interpolating untrusted input into its primary execution flow.
Ingestion points: User-provided task data enters the context via the {{ARGUMENTS}} placeholder in SKILL.md.
Boundary markers: Absent; the input is placed directly into the shell command and the task description without delimiters or instructions to ignore embedded commands.
Capability inventory: The skill has the capability to execute shell commands (omc ask) and write files to the .omc/artifacts/ directory.
Sanitization: None detected; the skill relies on the underlying platform or the omc tool to handle potentially malicious shell metacharacters or injected instructions.

ask