orca-per-workspace-env
Fail
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: HIGHPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The instructions explicitly direct the agent to re-run commands with escalated permissions if a 'runtime_access_denied' error is encountered, which is a directive to bypass security constraints.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from an external command output ('ORCA skills get') to define its operational logic. Ingestion points: The output of the
ORCA skills get orca-per-workspace-envcommand in SKILL.md. Boundary markers: None; the agent is told to load the guide directly. Capability inventory: The skill uses subprocess calls to execute various Orca CLI binaries and potentially interacts with cloud sandboxes or VMs. Sanitization: No validation or sanitization is mentioned for the content loaded from the tool output. - [COMMAND_EXECUTION]: The skill provides a logic flow for the agent to select and execute local binaries ('orca-dev', 'orca-ide', 'orca') based on the execution environment variables like 'ORCA_CLI_COMMAND' or 'ORCA_DEV_REPO_ROOT'.
Recommendations
- AI detected serious security threats
Audit Metadata