- Comience ya
- Mejores prácticas
- Capacidades
- Automation discovery (Preview)
- Artifact review (Preview)
- Operar
- Solución de problemas
- Aptitudes
- CLI
- Ejemplos
- Resumen de ejemplos
- Receta: crear un agente
- Receta: automatizar con RPA
- Receta: introducir datos en una aplicación web
- Receta: extraer datos de una aplicación de escritorio
- Receta: refactorizar y probar un flujo de trabajo
- Receta: un proceso basado en colas con REFramework
- Receta: crear un Maestro Flow
- Recipe: build a Maestro Case
- Receta: aplicación codificada y flujo de trabajo de API
- Recipe: build a coded function
- Receta: extraer datos de documentos con IXP
- Avanzado
- Ayuda
A read-only structural and quality audit of an RPA workflow, agent, flow, BPMN process, coded app, or solution, with findings classified by severity and routed to the skill that can fix them.
Point a coding agent at an existing UiPath project and it reviews it for structural validity, quality, best practices, and business alignment — without changing a single file. You get a findings report you can act on yourself, or hand back to the agent to fix.
Lo que hace
The agent reviews RPA workflows (.xaml, .cs), agents (.py, agent.json), Maestro Flows (.flow), BPMN processes (.bpmn), coded apps, and solutions (.uipx). It runs the artifact's own validation and build commands first, then evaluates structure, naming, error handling, and — where a process design document is available — whether the implementation matches the intended business logic.
Findings are classified by severity:
- Critical — blocks deployment.
- Warning — should be fixed.
- Info — an improvement opportunity, not a defect.
The review is read-only: the agent reports what it finds and recommends fixes, but the fix itself goes to the skill that builds that artifact type — for example, a workflow issue routes to uipath-rpa, an agent issue routes to uipath-agents.
La habilidad detrás de esto
This work is powered by the uipath-review skill, which discovers and classifies every project in scope, runs the relevant validation commands, and produces a structured findings report. You can see it in the Skills catalog.
Una solicitud de inicio típica
Una primera solicitud podría tener este aspecto:
"Revise esta automatización para detectar problemas de calidad y estructurales".
The agent classifies what it's reviewing, runs validation, and returns a findings report grouped by severity.
Más solicitudes para probar
- Review this agent project against its process design document and flag anything that doesn't match the intended business logic.
- Audit this solution's 12 projects and give me a summary plus a deep dive on the three riskiest.
- Check this workflow for security issues and incomplete error handling. Don't fix anything — just report.