The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute local binaries including orca, orca-ide, and orca-dev. It also explicitly directs the agent to prioritize the executable path defined in the ORCA_CLI_COMMAND environment variable, which could allow a user or environment to redirect the agent to a malicious binary.
[DYNAMIC_EXECUTION]: The skill uses a 'discovery stub' pattern where it commands the agent to run ORCA skills get orca-cli to fetch the 'complete, version-matched guide'. This results in the agent loading and following instructions that are generated at runtime by a local binary, bypassing static analysis of the skill file itself.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by ingesting and acting upon instructions provided by the output of the orca binary.
Ingestion points: Output of ORCA skills get orca-cli and ORCA status --json (SKILL.md).
Boundary markers: None provided; the agent is instructed to read the output and 'run the specific command you need'.
Capability inventory: Execution of subcommands including worktree, terminal, browser, and handoff (SKILL.md).
Sanitization: None; the agent is told to report exact errors and follow the guide's instructions.

orca-cli