orca-per-workspace-env
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides logic for resolving and executing a local command-line interface (CLI) binary (such as
orca,orca-dev, ororca-ide) based on environment variables (ORCA_CLI_COMMAND,ORCA_DEV_REPO_ROOT) or operating system context. It contains defensive instructions to prevent accidental execution of the GNOME Orca screen reader on Linux systems. - [DYNAMIC_EXECUTION]: The skill instructs the agent to dynamically load its full instruction set by executing
ORCA skills get orca-per-workspace-env. This ensures that the agent's documentation remains in sync with the local binary version, though it results in the agent's extended behavior being defined by the output of a local executable. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to interact with and validate user-controlled configuration files (
orca.yaml) and repository paths. - Ingestion points: The
orca vm recipe doctorcommand readsorca.yamland repository data. - Boundary markers: No explicit instructions are provided for the agent to treat these external configuration values as untrusted or to wrap them in specific delimiters.
- Capability inventory: The skill provides capabilities for local command execution and potential cloud resource provisioning (gated by user approval).
- Sanitization: The skill relies on the external binary's internal validation logic to handle the configuration data safely.
Audit Metadata