orca-cli
Warn
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to interact with the system via the
orcaCLI utility. It instructs the agent to execute various subcommands, includingstatus,worktree ps, andterminal list, to manage worktrees, terminals, and an embedded browser. - [REMOTE_CODE_EXECUTION]: The skill dynamically resolves the path to the Orca executable by checking the
ORCA_CLI_COMMANDenvironment variable and fallbacks likeorca-devororca-ide. Executing binaries based on untrusted environment variables can lead to the execution of arbitrary code if an attacker can manipulate the environment. - [PROMPT_INJECTION]: The skill uses a 'living document' pattern where it instructs the agent to run
ORCA skills get orca-clito fetch its primary usage guide. This creates an indirect prompt injection surface because the agent's behavior and constraints are determined at runtime by the output of an external tool, which lacks explicit boundary markers or sanitization in the provided instructions.
Audit Metadata