fix(orchestration,loop): 子代理现在真的收到任务(P0)+ 工具白名单 + 超时可配(裁决 02.4.2/02.4.1/02.4.4)
CI / Backend (build / vet / test / cross-compile) (push) Has been cancelled
CI / Frontend (tsc + vite build) (push) Has been cancelled

执行方案阶段 2 · P2-1。**02.4.2 是最伤的一条 P0**:`spawntool` 从不填 `AgentSpec.Context`、
`scheduler` 把 nil 交给 loop ⇒ 子代理模型上下文 = 角色 system + 记忆召回 + **0 条消息**,
goal/attach_info/data 一个字都到不了模型(此前测试用 mock adapter 恒答固定文本,测不出)。

- **02.4.2**:新增 `taskContext/taskUserMessage`(`spec.go`)把 **goal 正文 + attach_info + data
  渲染成同一条 user 消息**(`[任务目标]/[附加信息]/[任务数据]` 独立段落);`Spawn()` 设
  `Context: taskContext(spec)`;**system 只保留角色指令**,任务绝不进 system。
  真断言(捕获请求体的 adapter + 真 loop/agent):首请求恰好 **1 条 user 消息**、同时含 goal 与
  attach_info,且**所有 system 消息不含 goal 文本**。
- **02.4.1**:`AgentSpec.ToolWhitelist`(nil/空 = 继承全量);schema 增 `tool_whitelist`
  (顶层 + 每 task 条目,task 未带时继承批次级);`scheduler.run()` 经 ctx 传递 +
  `scopedRunner` 派生;`loop/options.go` **纯增量(+75 / −0)**:`WithToolWhitelist`/
  `ToolWhitelistFrom` + `(*Loop).WithToolWhitelist` 派生副本(浅拷贝仅替换 `hooks.CallAdmit`,
  链式保留原钩子),名单外调用否决为**软结果**回灌模型。
- **02.4.4**:`SUBAGENT_TIMEOUT_SEC`(默认 600)入 `internal/config` + `.env.example`;
  优先级 = 按调用 `timeout_seconds`(1s–24h 校验)> `WithSpawnTimeout`(生产由 cfg 填充)>
  调度器 `PerAgentTimeout`(`wire*.go` 仍是 10min 常量,未动);迟到 span 续根由测试锁定。

验证:**全树 `go build`/`go vet`/`go test ./... -count=1` = 58 包 ok / 0 FAIL**;
`GOOS=linux/amd64` + `GOOS=darwin/arm64` 交叉编译绿;前端 `tsc --noEmit` + `npm run build` 绿;
本批文件全部 ≤500 行、函数 ≤80 行。
变异验证(四条,均带还原证明):① `Context` 改回 `cloneMessages(spec.Context)` →
「1 条 user 消息」断言 FAIL(`user messages = 0`,即 P0 症状);② `WithToolWhitelist` 原样返回 →
显式白名单测试 FAIL(`alpha executed 1 times despite not being whitelisted`);
③ `subCtx` 改回全局默认 → 3 个超时测试 FAIL;④ `foldSpanFinish` 丢掉 `SpanFinished` →
迟到 span 状态断言 FAIL。

⚠️ 已知缺口(**登记,未假装已解决**):① 无真实模型端到端(全部捕获请求体 mock);
② 白名单**只挡执行、不过滤工具广告面**(模型仍能看见未列工具,调用即被拒并拿到可读原因;
过滤需改 `loop/request.go`);③ **迟到 span 的"新一轮已开始"窗口未修**(`loop/flow.go` 的
`StartRound` 会整体替换 `graphs[session]` ⇒ 新轮开始后迟到的 span 仍落进新轮图;需按轮键控图,
属后续);④ per-role 超时覆盖未实现(需 persona 旋钮管道)。
This commit is contained in:
AskaEth committed 2026-10-04 11:30:39 +08:00
1 parent 6e7ec56f9a
commit 7e7f53bb41
15 files changed
+974 -50

No files matched your search

+10 -3
View File
@@ -11,8 +11,8 @@ import (
)
// SubAgent is a single spawned sub-agent instance. The identity fields (ID, Goal,
// RoleID, PersonaID, Mode, AttachInfo, ParentSession, Data, Context, WorkingMemory)
// are set once at Spawn and never mutated. The lifecycle fields (Status, Summary,
// RoleID, PersonaID, Mode, AttachInfo, ParentSession, Data, ToolWhitelist,
// Context, WorkingMemory) are set once at Spawn and never mutated. The lifecycle fields (Status, Summary,
// Result, Err) are written by the sub-agent's background goroutine and read through
// the synchronized accessors below, so reading them before the sub-agent finishes is
// safe. A sub-agent is terminal exactly once its Done reports true; Join returns
@@ -40,12 +40,19 @@ type SubAgent struct {
ParentSession string
// Data is the optional structured payload supplied at spawn.
Data any
// ToolWhitelist 是该子代理的工具白名单(design/05 §7 裁决 02.4.1):
// nil/空=继承主 agent 全量工具集;非空=只放行所列工具。Spawn 时深拷贝。
ToolWhitelist []string
// Context is the isolated conversation context. It is a deep copy of the
// AgentSpec.Context and is never shared with the front session.
// AgentSpec.Context plus the task user message, and is never shared with the
// front session.
Context []model.Message
// WorkingMemory is the sub-agent's own working-memory window, created fresh
// at Spawn so it never overlaps another sub-agent or the front session.
WorkingMemory *memory.WorkingMemory
// timeout 是该子代理生效的运行时限(按调用覆盖优先于调度器默认,
// design/05 §7 裁决 02.4.4);非正表示除取消外无期限。Spawn 时落定,只读。
timeout time.Duration
mu sync.RWMutex
status Status