Replace assistant-prose heuristic with explicit turn lifecycle markers

Extract provider-authored turn boundaries (completion, interruption) directly
from Claude/Codex transcripts so the chat view knows when work ends without
guessing from message presence. Reconciles live hook state with transcript
lifecycle: when a terminal boundary lands, it settles a dropped Stop hook
instead of letting prose mislead the UI into showing 'working' after done.

src/main/native-chat/transcript-reader.ts

@@ -1,5 +1,9 @@
 import { createReadStream } from 'node:fs'
-import type { AgentType, NativeChatMessage } from '../../shared/native-chat-types'
+import type {
+  AgentType,
+  NativeChatMessage,
+  NativeChatTurnLifecycle
+} from '../../shared/native-chat-types'
 import { resolveNativeChatTranscriptAgent } from '../../shared/native-chat-agent-support'
 import { errorMessage } from '../ai-vault/session-scanner-values'
 import { resolveSessionFilePath, type ResolveSessionFileOptions } from './session-file-resolver'
@@ -11,7 +15,11 @@ import {
 import { decodeTranscriptStream } from './transcript-stream-lines'
 
 export type ReadTranscriptResult =
-  | { messages: NativeChatMessage[] }
+  | {
+      messages: NativeChatMessage[]
+      lifecycle?: NativeChatTurnLifecycle
+      turnLifecycleCapable?: boolean
+    }
   // notFound marks a retry-worthy miss (transcript not flushed to disk yet,
   // #8401) as opposed to a real parse/IO error callers surface immediately.
   | { error: string; notFound?: true }