orca-cli
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill uses a bootstrap mechanism where it fetches its primary instructions dynamically from the
orcabinary at runtime. - Ingestion points: The text content returned by running
ORCA skills get orca-cliin the terminal. - Boundary markers: None present; the agent is instructed to read the output and then execute commands based on it, treating the external output as authoritative instructions.
- Capability inventory: Shell command execution via the
orcabinary (status, open, worktree, terminal), which manages worktrees, repositories, and an embedded browser. - Sanitization: None; the agent is directed to treat the CLI output as the source of truth for subcommands and flags.- [COMMAND_EXECUTION]: The skill performs shell command execution using a resolved executable (
orca,orca-ide, ororca-dev). The resolution logic depends on environment variables likeORCA_CLI_COMMANDandORCA_DEV_REPO_ROOTto determine which binary to invoke.
Audit Metadata