opentrons-app

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the ocu (Open Computer Use) tool to automate desktop interactions and retrieve application status as described in the Core Workflow of SKILL.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes text from the Opentrons App, which can include content from user-provided protocol files.
  • Ingestion points: The get_app_state tool reads GUI text and accessibility metadata (SKILL.md).
  • Boundary markers: The skill relies on 'Safety boundary' instructions requiring user approval for consequential actions, but lacks technical delimiters to escape or isolate external text.
  • Capability inventory: The skill can execute commands to move robot hardware, calibrate instruments, and start laboratory protocols via UI automation.
  • Sanitization: No evidence of sanitization or character escaping for data extracted from the application interface.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 08:41 AM
Security Audit — agent-trust-hub — opentrons-app