sonnet

rumor-ml's avatarfrom rumor-ml

Override default model selection for specific subagents to use Sonnet instead of Opus. Use ONLY when the user explicitly requests to use Sonnet model for Plan, code-reviewer, code-simplifier, or Resolve Conflicts agents. Do not use this skill proactively - it requires explicit user instruction.

⭐1stars🔀0forks📁View on GitHub🕐Updated Dec 30, 2025
[Tool Routing and Orchestration]

When & Why to Use This Skill

The Sonnet Model Override skill allows users to optimize their Claude agent workflows by switching specific subagents—such as planning, code review, and conflict resolution—from the default Opus model to Sonnet. This skill is designed to enhance execution speed and reduce operational costs while maintaining high-quality output for standard development tasks through precise model orchestration.

Use Cases

  • Accelerating Code Reviews: Use the Sonnet model for routine code quality checks to receive faster feedback and optimize token usage.
  • Cost-Effective Planning: Switch the Plan agent to Sonnet for initial project brainstorming and task breakdown to save on costs without sacrificing logic.
  • Rapid Conflict Resolution: Resolve git merge conflicts more efficiently by leveraging Sonnet's speed for straightforward code merging tasks.
  • Efficient Code Simplification: Explicitly request Sonnet for the code-simplifier subagent to maintain a budget-friendly development cycle during refactoring.
namesonnet
descriptionOverride default model selection for specific subagents to use Sonnet instead of Opus. Use ONLY when the user explicitly requests to use Sonnet model for Plan, code-reviewer, code-simplifier, or Resolve Conflicts agents. Do not use this skill proactively - it requires explicit user instruction.

Sonnet Model Override

Overview

Use Sonnet model for specific subagents that normally run with Opus. This provides faster execution with lower cost while maintaining high quality for most tasks.

When to Use This Skill

IMPORTANT: Only use this skill when the user explicitly requests to use Sonnet. Examples:

  • "Use Sonnet for the code review"
  • "Run the plan agent with Sonnet"
  • "Use Sonnet model for this"

Do NOT use this skill proactively or by default.

Supported Subagents

This skill applies to the following subagents only:

  1. Plan - Planning and exploration agent
  2. pr-review-toolkit:code-reviewer - Code review agent
  3. pr-review-toolkit:code-simplifier - Code simplification agent
  4. Resolve Conflicts - Merge conflict resolution agent

Usage

When invoking any of the supported subagents, add the model parameter set to "sonnet":

Examples

Plan agent with Sonnet:

Task(
  subagent_type="Plan",
  model="sonnet",
  prompt="Plan the implementation of the new feature",
  description="Plan new feature"
)

Code reviewer with Sonnet:

Task(
  subagent_type="pr-review-toolkit:code-reviewer",
  model="sonnet",
  prompt="Review the recent changes for code quality issues",
  description="Review code changes"
)

Code simplifier with Sonnet:

Task(
  subagent_type="pr-review-toolkit:code-simplifier",
  model="sonnet",
  prompt="Simplify the code while preserving functionality",
  description="Simplify code"
)

Resolve Conflicts with Sonnet:

Task(
  subagent_type="Resolve Conflicts",
  model="sonnet",
  prompt="Resolve the merge conflicts in the current branch",
  description="Resolve merge conflicts"
)

Important Notes

  • The model parameter accepts: "sonnet", "opus", or "haiku"
  • Only add this parameter when explicitly requested by the user
  • Sonnet provides a good balance of speed, cost, and quality for most tasks
  • For particularly complex tasks, Opus may still be preferable (but only use if user requests it)