Support for Claude Models "Thinking Mode" and Custom Subagents in Kiro Desktop

[DevSissi]

Feature Description

Problem: Currently, Kiro Desktop lags behind the Kiro CLI in terms of advanced model capabilities. Specifically:

1. No "Thinking Mode": The Desktop version does not support Claude's thinking model (extended reasoning), whereas Kiro CLI already supports it.
2. No Custom Subagents: The Desktop version lacks the functionality to define or configure custom subagents.

Suggested Solution: Please bring feature parity to the Desktop version by:

• Directly enabling Claude's Thinking Model in the Desktop chat interface.
• Alternatively, implementing a native "Think Mode" tool within Kiro Desktop.
• Adding a UI for creating and managing Custom Subagents, similar to the flexibility offered in the CLI or other AI IDEs.

Use Case

Why this matters: This is a significant missed opportunity for user adoption and developer experience.

• Debugging & Transparency: (Critical) In complex scenarios, users need visibility into the model's "Chain of Thought" to understand the output. When the model generates an incorrect answer, being able to view the reasoning steps is crucial for users to pinpoint exactly where the logic failed and correct it.
• Complex Reasoning: For difficult coding tasks, users rely on the "Thinking Mode" to reduce hallucinations and improve logic. Without this, the Desktop experience feels limited compared to the CLI.
• Market Adoption: Many users are looking to migrate to Kiro from other AI IDEs. However, the lack of these features in the Desktop GUI is a major blocker.
• Consistency: Since the underlying capabilities exist in Kiro CLI, exposing them in the Desktop UI is essential for a unified ecosystem.

Updating these features would significantly increase user retention and attract developers looking for a more capable AI editor.

Additional Context

No response