feat(task-tool): Add model parameter for subagent LLM routing

[ziuus]

Summary

• Add optional model parameter to Task tool for specifying which LLM subagents should use
• Support both modelID and providerID/modelID formats
• Model parameter overrides agent default and parent message model
• Enables routing tasks to specialized LLMs (vision models, research-optimized models, etc.)

Changes

• task.ts: Added model to Parameters schema, implemented model selection logic (params.model > agent.default > parent message)
• task.txt: Updated documentation with new parameter and usage examples
• task.test.ts: Added 3 test cases for model parameter handling

Use Cases

// Vision model for image analysis
Task(description="Analyze screenshot", prompt="...", subagent_type="general", model="gpt-4-vision-preview")

// Research-optimized model
Task(description="Research topic", prompt="...", subagent_type="general", model="openai/gpt-4-turbo")

// Cost-efficient model for simple tasks
Task(description="Batch process", prompt="...", subagent_type="general", model="llama-3.1-8b")

Testing

• Added tests for model parameter parsing (with and without provider prefix)
• Test for fallback behavior when model not specified
• All existing tests continue to pass

Breaking Changes

None - model parameter is optional and backward compatible.

Closes #ZZZ (feature request for subagent model routing)

[github-actions]

This PR doesn't fully meet our contributing guidelines and PR template.

What needs to be fixed:

• PR description is missing required template sections. Please use the PR template.

Please edit this PR description to address the above within 2 hours, or it will be automatically closed.

If you believe this was flagged incorrectly, please let a maintainer know.

[github-actions]

The following comment was made by an LLM, it may be inaccurate:

Potential Duplicate PRs Found

Based on the search results, there are several related PRs that address similar functionality:

1. PR feat(task): add model override for subagents #18528 - feat(task): add model override for subagents

   • Directly related: This PR already implements model override functionality for subagents, which is very similar to the current PR's goal.

2. PR feat: add assign-model feature for subagent model selection #17570 - feat: add assign-model feature for subagent model selection

   • Related: Implements subagent model selection, which overlaps with the model parameter routing feature.

3. PR feat(agent): implement model tier selection with variant support for subagents #11377 - feat(agent): implement model tier selection with variant support for subagents

   • Related: Addresses model tier selection for subagents, similar conceptual approach.

4. PR feat(task): pass variant to subagent so it inherits parent's thinking level #12567 - feat(task): pass variant to subagent so it inherits parent's thinking level

   • Related: Involves passing model-related parameters through the task tool to subagents.

Recommendation: Check PR #18528 in particular, as it appears to be the most directly related existing implementation of model override for subagents.

[ziuus]

Closing in favor of #18528 which has a more complete implementation with permission gating (model_override). Will contribute to that PR instead.