The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides specific instructions for the agent to resolve the orca executable path based on environment variables or system context. It directs the agent to execute shell commands such as ORCA skills get orca-cli, ORCA status --json, and ORCA open --json to interact with the local application.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to load its full operational guide dynamically from the output of a CLI command. This creates an ingestion surface where the agent's behavior is guided by external (though local) data.
Ingestion points: The output of the command ORCA skills get orca-cli is loaded into the agent's context in SKILL.md.
Boundary markers: Absent; the skill directs the agent to "Load the full guide before running Orca commands" without specifying delimiters for the incoming text.
Capability inventory: The skill enables the agent to manage worktrees, folder contexts, terminals, and an embedded browser through the ORCA binary.
Sanitization: Absent; the agent is expected to read the raw output of the command to determine its next steps.

orca-cli