feat(session,runtime): 多角色并存—按会话选角（迁移0011 + 按回合解析 + agent池 + 提取器修正）

按 DevDocs/design/22 批次 1 后端（§2/§3/§4.1-4.5/§4.7/§5 验收 1-6）。
- 迁移 0011：chat_sessions.persona_id + chat_messages.persona_id（均 DEFAULT ''，无回填 → 升级零风险）；同步包内幂等 DDL；**真实库升级验证**（拷贝 data/cyrene.db 到临时目录跑 MigrateUp：既有 1 会话/109 消息 persona 仍为 ''，原库未被触碰）
- history：SessionMeta/Message 加 PersonaID + SetSessionPersona；ListSessions/List/scan/Append 带上
- runtime：SessionEnv.PersonaID + loop.personaFor(ctx) 替换固定 l.personaID（memory/request/turn）+ **agent 池**（Options.Agents，按角色取 instructions，未命中回落）+ **提取器不再把 persona 烤死在构造期**（多角色下记忆不再永远记在 cyrene 名下）+ **thinking 旋钮改为按会话角色解析**（否则角色 Y 会用角色 X 的预算/展示）
- gateway：会话 DTO/POST/PATCH 支持 personaId；ws/rest/voice 的 OwnerEnv 硬编改为**按会话解析角色**（actor 元数据语义不变）；admin memories 搜索去掉硬编 cyrene
- Docs/api-client v0.42；**两处有意的行为收紧**已写明：POST /sessions body 非法 JSON → 400（原先完全忽略 body；空/null/{} 仍 201）；POST /admin/memories/search 省略 personaId 且无默认角色 → 400（原先默认查 cyrene）
- 测试：design/22 §5 验收 1-6（记忆隔离/提示词首条 system/情绪按角色/提取 persona 归属/迁移兼容）+ 进程内 e2e（断言 adapter 收到的 model.Request.Messages[0] 即该角色的 systemPrompt）
- **已知未做（留批次 2 / 待确认）**：① IM 与渠道级未做 → **§7-3「owner 私聊镜像随主会话角色」尚未生效**；② 请求级覆盖未实现（本批只做会话级）；③ **§7-5 的「上下文发送角色标记」只完成存储+DTO 半边**——消息 persona 已落库且 DTO 已带，但模型上下文不从 chat_messages 重建，「第 N 条是谁说的」尚未进上下文（需确认是否属批次 2）；④「切角色重建摘要」当前是空操作（压缩未接线）；⑤ thinking 旋钮仅在流式路径生效（非流式 agent.RunStep 不带，A 批次既有性质）
- 验收：go build/vet/test 全绿 + linux/amd64 与 darwin/arm64 交叉编译（既有 flake 本轮未触发；同批 session_orchestrator.go 一度 501 行 → 抽出 session_tail.go）

internal/runtime/loop/stream.go

@@ -15,11 +15,16 @@ import (
 // path, this method is owned by the loop; it mirrors the agent's semantics so the
 // two behave identically apart from the delta callback.
 func (l *Loop) streamStep(ctx context.Context, messages []model.Message) (*stepOutcome, error) {
-	adapter := l.agent.Adapter()
+	ag := l.agentFor(ctx)
+	adapter := ag.Adapter()
 	if adapter == nil {
 		return nil, fmt.Errorf("loop: stream step: %w", ErrNoAdapter)
 	}
-	req := l.buildRequest(ctx, messages)
+	// Resolve the role's agent and reasoning knobs once per step: the same pair is
+	// used for the request and for the per-delta broadcast gate, so a role switch
+	// can never apply the new prompt with the old broadcast switch mid-stream.
+	knobs := l.knobsFor(ctx)
+	req := l.buildRequest(ctx, messages, ag, knobs)
 
 	if l.guards != nil {
 		verdict, err := l.guards.CheckInput(ctx, &req)
@@ -40,10 +45,10 @@ func (l *Loop) streamStep(ctx context.Context, messages []model.Message) (*stepO
 		ev.TurnID = loopTurn(ctx)
 		// Suppress UI broadcast for internal autonomous passes (backthink),
 		// whose intermediate reasoning must not leak into the chat stream, and for
-		// thinking frames when the bound persona turned them off (showToUser:
-		// false). Both gates are broadcast-only: the response the model produced is
+		// thinking frames when the turn's role turned them off (showToUser: false).
+		// Both gates are broadcast-only: the response the model produced is
 		// unchanged, so display and persistence stay decoupled.
-		if l.onStream != nil && !isMutedBroadcast(ctx) && l.broadcasts(ev) {
+		if l.onStream != nil && !isMutedBroadcast(ctx) && broadcasts(ev, knobs.ThinkingShowToUser) {
 			l.onStream(ev)
 		}
 	})
@@ -70,15 +75,15 @@ func (l *Loop) streamStep(ctx context.Context, messages []model.Message) (*stepO
 }
 
 // broadcasts reports whether a streamed event may be pushed to the UI. Only
-// thinking frames are ever withheld, and only when the wiring resolved the
-// persona's showToUser to false; text, tool-call and tool-progress frames are
-// never affected, so turning thinking off cannot change what the client sees of
-// the answer itself.
-func (l *Loop) broadcasts(ev model.StreamEvent) bool {
+// thinking frames are ever withheld, and only when the turn's role resolved
+// showToUser to false; text, tool-call and tool-progress frames are never
+// affected, so turning thinking off cannot change what the client sees of the
+// answer itself.
+func broadcasts(ev model.StreamEvent, showToUser *bool) bool {
 	if !isThinkingKind(ev.Kind) {
 		return true
 	}
-	return l.thinkingShow == nil || *l.thinkingShow
+	return showToUser == nil || *showToUser
 }
 
 // isThinkingKind reports whether kind is one of the reasoning segment events.