The Agent Skills Directory

[COMMAND_EXECUTION]: The skill resolves and executes a local command-line utility (orca, orca-ide, or orca-dev) to perform its core functions. The specific executable path is determined at runtime by checking environment variables such as ORCA_CLI_COMMAND and ORCA_DEV_REPO_ROOT.
[DYNAMIC_EXECUTION]: The skill uses a "discovery stub" pattern where it dynamically fetches its functional instructions by running ORCA skills get orca-cli. This results in the agent's operational logic being defined by the output of a shell command executed during the session.
[INDIRECT_PROMPT_INJECTION]: By design, the skill processes instructions provided by an external local process (the CLI tool). This creates an attack surface where a compromised binary or environment could inject malicious instructions into the agent's context.
Ingestion points: The output of the ORCA skills get orca-cli command is directly loaded as a guide for the agent in the SKILL.md file.
Boundary markers: No explicit delimiters or safety warnings are present to distinguish the fetched instructions from the agent's core safety guidelines.
Capability inventory: The skill provides capabilities for worktree management, terminal interaction, repository operations, and browser control via the orca CLI.
Sanitization: There is no evidence of sanitization or validation of the content returned by the CLI before it is incorporated into the agent's context.

orca-cli