refactor(models): Refine MessageAgentThought SQLAlchemy typing

[tomerqodo]

Benchmark PR from agentic-review-benchmarks#11

[sentry]

Bug: The organize_agent_history method incorrectly uses tool_inputs instead of tool_responses for the content of a ToolPromptMessage, breaking the agent's reasoning.
_{Severity: CRITICAL}

Suggested Fix

In api/core/agent/base_agent_runner.py on line 502, change content=str(tool_inputs.get(tool, agent_thought.observation)) to content=str(tool_responses.get(tool, agent_thought.observation)) to correctly use the tool's response content.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent.
Verify if this is a real issue. If it is, propose a fix; if not, explain why it's not
valid.

Location: api/core/agent/base_agent_runner.py#L502

Potential issue: In the `organize_agent_history` method, when constructing the
conversation history, the content for the `ToolPromptMessage` is incorrectly sourced
from the `tool_inputs` dictionary instead of the `tool_responses` dictionary. This
causes the agent to receive its own tool call inputs as if they were the tool's output.
This breaks the logical flow of the conversation history, as the
`AssistantPromptMessage` contains the tool call inputs, and the subsequent
`ToolPromptMessage` should contain the corresponding responses. This will prevent the
agent from correctly reasoning about the results of its tool executions.

_{Did we get this right? 👍 / 👎 to inform future reviews.}