feat(agent): 支持为每个 agent 绑定 LLM 配置(agent 级默认模型)

按优先级 Agent 绑定 > 任务/会话指定 > 全局激活 解析每个 agent 运行的模型,
使 planner/worker/goals/mainagent/自定义 agent 可各自走不同 profile(如 plan 用
强模型、work 用弱模型)。未绑定时行为与原先完全一致。

- db: Agent 增 llm_profile_id + SetAgentLLMProfile;schema 幂等兜底 ALTER
- server: 新增 effectiveProfileForAgent/providerForProfile/providerForAgent;
  buildPlannerWorker 让 planner/worker 各自解析 provider;provider 按 profile id
  缓存以共享限速器;goals/会话链同步走绑定;绑定变更热重载
- api: agent 配置接口收 llm_profile_id(三态) + 返回可选 profile 列表
- web: agent 编辑器「配置与提示词」页新增「默认模型」下拉
This commit is contained in:
Autumn-27 committed 2026-08-11 12:15:38 -04:00
1 parent 289cf6a24f
commit c84bba477f
10 files changed
+238 -55

No files matched your search

+9 -4
View File
@@ -62,11 +62,16 @@ func (s *Server) createGoals(ctx context.Context, t *Task, emit func(db.Activity
return specs
}
// goalsLLMConfig resolves the LLM config for goal decomposition: the task's pinned
// profile if any, else the global active profile (same source the engine runs on).
// goalsLLMConfig resolves the LLM config for goal decomposition by the standard
// precedence: the goals agent's own binding → the task's pinned profile → the global
// active profile (same source the engine runs on).
func (s *Server) goalsLLMConfig(t *Task) (agent.Config, bool) {
if t != nil && t.LLMProfileID != nil {
if cfg, ok := s.loadProfileConfig(*t.LLMProfileID); ok {
var pin *int64
if t != nil {
pin = t.LLMProfileID
}
if eff := s.effectiveProfileForAgent("goals", pin); eff != nil {
if cfg, ok := s.loadProfileConfig(*eff); ok {
return cfg, true
}
}