feat(inference): capture reasoning_content on responses (WIP, provider side)

Adds a nullable `reasoning` field to InferenceResponse and maps the
llama.cpp / OpenAI-compat `reasoning_content` field into it, so local models
that emit their chain-of-thought have it captured. Provider/model side only —
the emit-site event field (reasoningArtifactId) + CAS storage in the
orchestrator are still pending (tracked in Vikunja Correx #4).

core/inference/src/main/kotlin/com/correx/core/inference/InferenceResponse.kt

@@ -11,4 +11,8 @@ data class InferenceResponse(
     val tokensUsed: TokenUsage,
     val latencyMs: Long,
     val toolCalls: List<ToolCallRequest> = emptyList(),
+    // The model's chain-of-thought, when the provider exposes it separately (llama.cpp / OpenAI-compat
+    // `reasoning_content`). Captured for observability/audit — many local models emit their whole plan
+    // here and leave `text` empty, so dropping it hid why a stage behaved as it did (2026-07-05).
+    val reasoning: String? = null,
 )