Fix OpenAI reasoning item error in Responses API (#59)

## Problem

OpenAI Responses API was intermittently failing with error:
```
Item 'rs_00d363a611783a350068e4764c5f68819c8777140c3248eff4' of type 'reasoning' was provided without its required following item.
```

This occurred when using reasoning models (gpt-5, o3, o4-mini) with tool
calls in multi-turn conversations.

## Root Cause

From the [Vercel AI SDK
documentation](https://sdk.vercel.ai/providers/ai-sdk-providers/openai#responses-models):

> When using reasoning models (o1, o3, o4-mini) with multi-step tool
calls and `store: false`, include `['reasoning.encrypted_content']` in
the `include` option to ensure reasoning content is available across
conversation steps.

Even though we're using the default `store: true` and
`previousResponseId` for persistence, we still need to explicitly
include reasoning encrypted content when tool calls are involved. The
reasoning items have IDs (like `rs_*`) that must be properly linked to
their following items.

## Solution

Added `include: ['reasoning.encrypted_content']` to OpenAI provider
options when `reasoningEffort` is configured (meaning reasoning is
enabled).

This ensures reasoning context is properly preserved across multi-turn
conversations with tool calls.

## Changes

- Updated `buildProviderOptions()` in `src/utils/ai/providerOptions.ts`
- Only adds `include` when reasoning is actually enabled
- Added comments explaining why this is required

## Testing

Not adding automated tests as the error was intermittent and difficult
to reliably reproduce. The fix is based directly on OpenAI/Vercel SDK
documentation requirements.

Author: ammario

src/utils/ai/providerOptions.ts

@@ -115,6 +115,10 @@ export function buildProviderOptions(
         ...(reasoningEffort && {
           reasoningEffort,
           reasoningSummary: "detailed", // Enable detailed reasoning summaries
+          // Include reasoning encrypted content to preserve reasoning context across conversation steps
+          // Required when using reasoning models (gpt-5, o3, o4-mini) with tool calls
+          // See: https://sdk.vercel.ai/providers/ai-sdk-providers/openai#responses-models
+          include: ["reasoning.encrypted_content"],
         }),
         // Include previousResponseId for persistence (Responses API)
         ...(previousResponseId && { previousResponseId }),