uipath-maestro-bpmn

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The skill explicitly instructs agents to overwrite failed CLI-discovery JSON with a local validator file (fabricating expected registry-evidence) instead of preserving real discovery output, which is a deceptive instruction outside the stated purpose of saving raw registry evidence.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In the operate/diagnose runtime path, the skill ingests outsider-authored free text from deployed BPMN runtime artifacts returned by UiPath CLI reads (e.g., instance incidents/variables/element executions and deployed asset via uip maestro bpmn instance ... --output json), where those records can contain user-supplied payloads/messages; this data is then parsed and used to generate diagnosis output.