Add stopReason to AgentLoop result when stop() is called

[chinmaymk]

Summary

This PR adds support for capturing and returning a stop reason when the agent loop is stopped via ctx.stop(reason). The stop reason is now included in the LoopResult as an optional stopReason field.

Key Changes

• Updated StoppableContext.stop() signature to accept an optional reason parameter in src/agent/types.ts
• Enhanced LoopResult interface to include an optional stopReason?: string field in src/agent/loop.ts
• Modified stop function implementation to capture the provided reason and log it when stopping the loop
• Updated return statements to conditionally include stopReason in the result object only when a reason was provided
• Added comprehensive test coverage with two new test cases:
  • Verifies that ctx.stop(reason) includes the stop reason in the result
  • Verifies that stopReason is undefined when stop() is called without a reason

Implementation Details

• The stopReason is stored in a closure variable and conditionally spread into the return object using the pattern ...(stopReason && { stopReason })
• A debug log statement is included when a stop reason is provided
• The feature is backward compatible - existing code calling stop() without arguments continues to work as before

https://claude.ai/code/session_017FisrUENRPyYjwgErHWAYt