orca-cli
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands using the orca, orca-ide, or orca-dev binaries. The specific executable path can be overridden by the ORCA_CLI_COMMAND environment variable.\n- [DYNAMIC_EXECUTION]: The skill fetches its full instruction set at runtime using the command 'ORCA skills get orca-cli'. This allows the external binary to dynamically inject instructions into the agent's context, bypassing static review of the skill's instructions.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes data from terminals and worktree statuses which can contain untrusted content.\n
- Ingestion points: Terminal lists, worktree process lists, and status commands (ORCA status, ORCA worktree ps, ORCA terminal list).\n
- Boundary markers: Absent. The agent is not instructed to treat the output of these commands as untrusted data.\n
- Capability inventory: The skill can execute subcommands, control an embedded browser, and manage file systems via worktrees.\n
- Sanitization: Absent. There is no evidence of filtering or validation of the CLI output before it is processed by the agent.
Audit Metadata