computer-use
Warn
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: MEDIUMPRIVILEGE_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_deniederror occurs. This encourages the agent to seek or use higher privilege levels to bypass system access controls. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates GUI automation of native applications and external browser windows (Chrome, Edge, Safari). This creates a vulnerability surface where the agent processes untrusted data from these windows which could contain malicious instructions.
- Ingestion points: External application windows, accessibility trees, and browser content via
orca computer(SKILL.md). - Boundary markers: None identified in the discovery stub.
- Capability inventory: GUI control including clicking, typing, and menu interaction; shell command execution via the
orcaCLI (SKILL.md). - Sanitization: No explicit sanitization or filtering of GUI content is described.
- [COMMAND_EXECUTION]: The skill relies on executing the
orcacommand-line interface and its variants (orca-ide,orca-dev) to perform system operations. These tools are identified as vendor resources associated with the skill's functionality.
Audit Metadata