The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill implements a dynamic instruction loading pattern by directing the agent to run ORCA skills get orca-cli to retrieve the version-matched usage guide. This output is treated as the source of truth for subsequent agent actions.
Ingestion points: The output of the ORCA skills get orca-cli command is ingested into the agent context (SKILL.md).
Boundary markers: The skill lacks explicit instructions for the agent to treat the binary output as untrusted data or to ignore embedded instructions.
Capability inventory: The resolved Orca binary has capabilities to operate worktrees, terminals, repositories, automations, and an embedded browser (SKILL.md).
Sanitization: There is no evidence of sanitization or filtering of the content returned by the binary before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill uses a resolution logic based on environment variables (ORCA_CLI_COMMAND, ORCA_DEV_REPO_ROOT) and system state to determine which executable (orca-dev, orca-ide, or orca) to run. While standard for CLI discovery, this involves executing commands from computed paths.
[COMMAND_EXECUTION]: The skill facilitates shell command execution by instructing the agent to resolve a local binary and execute subcommands (e.g., status, open, worktree ps) to manage application state.

orca-cli