orca-cli
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to load its operational guide at runtime by executing
ORCA skills get orca-cli. This design introduces a surface where the agent ingests external content into its instruction set. - Ingestion points: The content returned by the
ORCA skills getcommand is used as a "version-matched guide" for the agent. - Boundary markers: The skill does not define any delimiters or instructions to ignore potential malicious prompts embedded within the loaded guide.
- Capability inventory: The
orcaCLI has extensive capabilities, including managing worktrees, terminals, and an embedded browser, which could be abused if the guide content is compromised. - Sanitization: There is no mention of validation or filtering for the content returned by the CLI tool.
- [COMMAND_EXECUTION]: The skill instructs the agent to resolve and execute specific CLI binaries (
orca,orca-ide, ororca-dev). While this is the intended purpose of a CLI-wrapper skill, these commands interact with the host system's terminal, filesystem, and worktrees.
Audit Metadata