fix(core): ensure tool_choice maps to structured object for AWS Bedrock compliance

[toshalkumbhar8979-design]

Description

Ensures that tool_choice is formatted as a structured object ({"type": "auto"}) rather than a raw string when communicating via LiteLLM/Bedrock. This fixes an API validation failure where AWS Bedrock rejects requests with a 400 Bad Request.

Error Log

litellm.exceptions.BadRequestError: BedrockException - {"message":"The model returned the following errors: tool_choice.type: Field required"}

[greptile-apps]

Greptile Summary

This PR normalizes tool_choice before agent execution. The main changes are:

• Adds a pre-run conversion in strix/core/execution.py.
• Converts configured tool_choice values to a Bedrock-style object.
• Applies the conversion inside _run_cycle before Runner.run_streamed.

Confidence Score: 4/5

The Bedrock request path still has cases that can send the raw string, and explicit tool selections can be changed to auto.

• extra_body == {} skips top-level tool_choice normalization.
• Any string tool choice is rewritten to automatic selection.
• The change is localized to the pre-run request setup.

strix/core/execution.py

Important Files Changed

Filename	Overview
strix/core/execution.py	Adds tool_choice normalization before streamed agent runs.

Prompt To Fix All With AI

Fix the following 2 code review issues. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 2
strix/core/execution.py:354-358
**Top-Level Tool Choice Skips Normalization**

When `extra_body` is a dict without its own `tool_choice`, this `elif` skips the top-level `run_config.tool_choice` branch. A Bedrock run with `extra_body == {}` and `tool_choice == "auto"` can still reach `Runner.run_streamed` with the raw string and fail with the same `tool_choice.type` validation error.

```suggestion
            if hasattr(run_config, "extra_body") and isinstance(run_config.extra_body, dict):
                if "tool_choice" in run_config.extra_body:
                    run_config.extra_body["tool_choice"] = {"type": "auto"}
            if hasattr(run_config, "tool_choice") and (run_config.tool_choice == "auto" or isinstance(run_config.tool_choice, str)):
                run_config.tool_choice = {"type": "auto"}
```

### Issue 2 of 2
strix/core/execution.py:357-358
**Explicit Tool Choice Becomes Auto**

The `isinstance(run_config.tool_choice, str)` branch rewrites every string value to `{"type": "auto"}`. If a caller selects a specific tool by name, the request is changed to automatic selection before `Runner.run_streamed`, so the model can skip the required tool or call a different one.

```suggestion
            elif hasattr(run_config, "tool_choice") and run_config.tool_choice == "auto":
                run_config.tool_choice = {"type": "auto"}
```

_{Reviews (1): Last reviewed commit: "fix(core): ensure tool_choice maps to st..." | Re-trigger Greptile}

[greptile-apps]

Explicit Tool Choice Becomes Auto

The isinstance(run_config.tool_choice, str) branch rewrites every string value to {"type": "auto"}. If a caller selects a specific tool by name, the request is changed to automatic selection before Runner.run_streamed, so the model can skip the required tool or call a different one.

Suggested change

	elif hasattr(run_config, "tool_choice") and (run_config.tool_choice == "auto" or isinstance(run_config.tool_choice, str)):
	run_config.tool_choice = {"type": "auto"}
	elif hasattr(run_config, "tool_choice") and run_config.tool_choice == "auto":
	run_config.tool_choice = {"type": "auto"}

Prompt To Fix With AI

This is a comment left during a code review.
Path: strix/core/execution.py
Line: 357-358

Comment:
**Explicit Tool Choice Becomes Auto**

The `isinstance(run_config.tool_choice, str)` branch rewrites every string value to `{"type": "auto"}`. If a caller selects a specific tool by name, the request is changed to automatic selection before `Runner.run_streamed`, so the model can skip the required tool or call a different one.

```suggestion
            elif hasattr(run_config, "tool_choice") and run_config.tool_choice == "auto":
                run_config.tool_choice = {"type": "auto"}
```

How can I resolve this? If you propose a fix, please make it concise.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!