voiceuse-computer-control
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions create a surface for indirect prompt injection by directing the agent to process and act upon information observed on the local desktop screen.
- Ingestion points: Desktop screen content ingested via the
voiceuse_observe_screentool as described inSKILL.md. - Boundary markers: No specific instructions or delimiters are provided to help the agent distinguish between legitimate user UI elements and potentially malicious instructions embedded in displayed documents or web pages.
- Capability inventory: The skill leverages tools for mouse clicks (
voiceuse_click), text input (voiceuse_type_text), key presses (voiceuse_press_key), and application control (voiceuse_open_app). - Sanitization: There are no instructions for the agent to sanitize or validate data extracted from the screen before using it to inform subsequent actions.
Audit Metadata