The Agent Skills Directory

[COMMAND_EXECUTION]: The skill resolves and executes the orca CLI or its variants (orca-ide, orca-dev) based on the presence of environment variables like ORCA_CLI_COMMAND and ORCA_DEV_REPO_ROOT, allowing the execution path to be determined by the environment configuration.
[DYNAMIC_EXECUTION]: The skill dynamically loads instructions at runtime by executing ORCA skills get orca-cli. This command fetches a "version-matched guide" which provides the actual instructions for operating the CLI, effectively allowing the tool's output to modify the agent's prompt during the session.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process data from potentially untrusted sources such as worktree comments, artifacts, and terminal buffers, which creates an attack surface for indirect instructions.
Ingestion points: Data is read from orca terminals, worktree comments, and artifact content.
Boundary markers: There are no explicit instructions or delimiters provided to isolate or sanitize the data retrieved from these sources.
Capability inventory: The skill can execute shell commands via the CLI, manage terminals, and control an embedded browser.
Sanitization: No specific sanitization or validation logic is described for the content read from worktrees or terminals.

orca-cli