[Rovo Dev] Add model selection and reasoning effort configuration for personal agents in Studio

[MariaKolesnikova]

When creating a personal agent in Rovo Studio, users have no ability to configure the underlying AI model or reasoning effort. These are fundamental settings that directly affect agent quality, cost, and suitability for different tasks.

Note: model switching exists in Rovo Dev CLI via /models command, but this capability is not available when creating or configuring agents in Studio.

---

Steps to Reproduce:

1. Open Studio from the app switcher
2. Select Agents → Create (or edit an existing personal agent)
3. Go through all available configuration fields: Identity, Scenarios, Knowledge, Skills
4. Observe that no model selector or reasoning effort control is present anywhere in the UI

---

Expected vs Actual Behavior:

	Expected	Actual
Model selection	User can choose the AI model powering the agent in Studio UI	No model selector present — only available in CLI via /models
Reasoning effort	User can set reasoning depth (e.g., low / medium / high) per agent	Not available in any Rovo agent configuration surface

---

Use Case / Business Justification:

Different agent use cases have fundamentally different requirements:

• Simple, repetitive tasks (e.g., code formatting, changelog generation) benefit from a faster, lighter model with low reasoning effort — reducing latency and credit consumption.
• Complex tasks (e.g., architectural analysis, multi-file refactoring, debugging) require a more capable model with high reasoning effort to produce reliable results.

Without these controls in Studio, users are forced into a one-size-fits-all configuration. This is especially inconsistent given that Rovo Dev CLI already exposes model selection — the gap creates a degraded experience for users who build and manage agents through Studio.

---

Proposed Solution:

Add model selector and reasoning effort controls to the agent creation/editing UI in Studio — either in the Identity section or as a dedicated "Model" configuration tab.

[Dr Valeri Colon _Connect Centric_]

@MariaKolesnikova welcome to the community. This is a common ask. Today, model selection and “reasoning effort” aren’t exposed in Studio—those are managed behind the scenes, while CLI has more flexibility. Rovo is designed to balance performance, cost, and safety automatically, which is why it feels like a one-size setup. For now, the main levers in Studio are instructions, scenarios, and knowledge, not the underlying model.