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）

cmd/cyrene/persona_prompt.go

@@ -4,6 +4,7 @@ import (
 	"strings"
 
 	"git.yeij.top/AskaEth/Cyrene_v2/internal/persona/loader"
+	"git.yeij.top/AskaEth/Cyrene_v2/internal/runtime/loop"
 	"git.yeij.top/AskaEth/Cyrene_v2/internal/runtime/model"
 )
 
@@ -106,7 +107,9 @@ func bulletList(items []string) string {
 
 // defaultThinkingOptions returns the reasoning knobs the loop should inject for
 // the default persona: the thinking directive, the reasoning_effort value and the
-// thinking-broadcast switch.
+// thinking-broadcast switch. It is the default half of the per-role resolution
+// the loop performs every turn (Options.PersonaKnobs, see personaKnobs below), so
+// the configured default and the per-turn lookup can never drift apart.
 //
 // The thinking directive is non-nil only when the persona configures something
 // that affects the request (a budget, an effort, or an explicit on/off mode).
@@ -115,12 +118,35 @@ func bulletList(items []string) string {
 // effect. All three are zero when no persona is registered, which leaves the
 // loop's pre-existing behavior untouched.
 func (a *app) defaultThinkingOptions() (*model.Thinking, string, *bool) {
-	p, ok := a.defaultPersona()
+	knobs, ok := a.personaKnobs(a.defaultPersonaID())
 	if !ok {
 		return nil, "", nil
 	}
+	return knobs.Thinking, knobs.ReasoningEffort, knobs.ThinkingShowToUser
+}
+
+// personaKnobs resolves a role's reasoning knobs for the loop
+// (loop.Options.PersonaKnobs, design/22 §4.2): the loop asks for the role of the
+// turn it is about to build a request for, so a session bound to role Y runs with
+// Y's thinking budget / effort / broadcast switch instead of the default role's.
+// Resolving from the role id rather than pushing a value into every caller's
+// SessionEnv means every path — Web REST/WS/voice, IM, backthink, sub-agents —
+// gets the right knobs from one input. ok=false means "unknown role", and the loop
+// keeps its configured defaults.
+func (a *app) personaKnobs(personaID string) (loop.LoopKnobs, bool) {
+	if personaID == "" {
+		return loop.LoopKnobs{}, false
+	}
+	p, ok := a.personas.Persona(personaID)
+	if !ok || p == nil {
+		return loop.LoopKnobs{}, false
+	}
 	thinking, effort := personaThinking(p)
-	return thinking, effort, p.Thinking.ShowToUser
+	return loop.LoopKnobs{
+		Thinking:           thinking,
+		ReasoningEffort:    effort,
+		ThinkingShowToUser: p.Thinking.ShowToUser,
+	}, true
 }
 
 // personaThinking maps the persona's `thinking` section onto the model-layer