feat(engine): 添加运行时上下文支持并重构工具迭代限制

添加 RuntimeContext 类用于捕获模型运行时的日期时间信息，
包括UTC时间、本地时间和时区信息，并在系统提示中显示这些信息。

同时增加最大上下文消息数和工具迭代次数的配置选项，
将验证服务从引擎加载器中移除，并更新相关的数据结构和接口。

BREAKING CHANGE: 移除了验证服务，相关字段被替换为证据状态和接受状态。

- 添加 RuntimeContext 类和相关渲染方法
- 增加 max_context_messages 和 max_tool_iterations 配置
- 移除 ValidationService 相关代码
- 更新消息记录中的验证状态字段
- 添加原始工具调用检测和回退处理

app-instance/backend/beaver/tasks/service.py

@@ -7,7 +7,7 @@ from pathlib import Path
 from typing import Any
 from uuid import uuid4
 
-from .models import TaskEvent, TaskRecord, ValidationResult
+from .models import TaskEvent, TaskRecord
 from .store import TaskStore
 
 
@@ -105,38 +105,70 @@ class TaskService:
         for name in skill_names or []:
             if name not in task.skill_names:
                 task.skill_names.append(name)
+        task.status = "awaiting_acceptance"
         task.updated_at = self._now()
         self.store.upsert_task(task)
         self._event(task, "run_completed", run_id=run_id, payload={"skill_names": skill_names or []})
+        self._event(task, "evidence_recorded", run_id=run_id, payload={"skill_names": skill_names or []})
         return task
 
-    def record_validation(
+    def add_acceptance(
         self,
         task_id: str,
-        run_id: str,
-        validation: ValidationResult,
         *,
-        final_attempt: bool = True,
-        has_usable_answer: bool = True,
+        acceptance_type: str,
+        comment: str | None = None,
+        run_id: str | None = None,
     ) -> TaskRecord:
         task = self._require(task_id)
         now = self._now()
-        if validation.status == "accepted":
-            task.status = "awaiting_feedback"
-        elif validation.status in {"insufficient_evidence", "validator_error"}:
-            task.status = "needs_review"
-        elif validation.status == "rejected" and not final_attempt:
+        normalized = normalize_acceptance_type(acceptance_type)
+        matching_acceptance = any(
+            item.get("run_id") == run_id and item.get("acceptance_type") == normalized
+            for item in task.feedback
+        )
+        conflicting_acceptance = next(
+            (
+                item
+                for item in task.feedback
+                if item.get("run_id") == run_id and item.get("acceptance_type") != normalized
+            ),
+            None,
+        )
+        if conflicting_acceptance is not None:
+            raise ValueError(
+                f"Acceptance for run_id={run_id!r} was already recorded as "
+                f"{conflicting_acceptance.get('acceptance_type')!r}"
+            )
+        if task.status in {"closed", "abandoned"} and not matching_acceptance:
+            raise ValueError(f"Task {task.task_id} is already finalized as {task.status!r}")
+        if matching_acceptance:
+            return task
+
+        entry = {
+            "acceptance_type": normalized,
+            "feedback_type": "satisfied" if normalized == "accept" else normalized,
+            "comment": comment or "",
+            "run_id": run_id,
+            "created_at": now,
+        }
+        task.feedback.append(entry)
+        if normalized == "revise":
             task.status = "needs_revision"
-        elif validation.status == "rejected" and has_usable_answer:
-            task.status = "needs_review"
-        else:
-            task.status = "failed"
+        elif normalized == "abandon":
+            task.status = "abandoned"
             task.closed_at = now
-            task.close_reason = "automatic validation rejected the final attempt"
+            task.close_reason = comment or "abandoned"
+        elif normalized == "accept":
+            task.status = "closed"
+            task.closed_at = now
+            task.close_reason = "accepted"
+            task.satisfaction = 1.0
+            if run_id:
+                task.metadata["final_accepted_run_id"] = run_id
         task.updated_at = now
-        task.validation_result = validation.to_dict()
         self.store.upsert_task(task)
-        self._event(task, "validated", run_id=run_id, payload=validation.to_dict())
+        self._event(task, f"acceptance_{normalized}", run_id=run_id, payload=entry)
         return task
 
     def add_feedback(
@@ -147,52 +179,12 @@ class TaskService:
         comment: str | None = None,
         run_id: str | None = None,
     ) -> TaskRecord:
-        task = self._require(task_id)
-        now = self._now()
-        matching_feedback = any(
-            item.get("run_id") == run_id and item.get("feedback_type") == feedback_type
-            for item in task.feedback
+        return self.add_acceptance(
+            task_id,
+            acceptance_type=feedback_type,
+            comment=comment,
+            run_id=run_id,
         )
-        conflicting_feedback = next(
-            (
-                item
-                for item in task.feedback
-                if item.get("run_id") == run_id and item.get("feedback_type") != feedback_type
-            ),
-            None,
-        )
-        if conflicting_feedback is not None:
-            raise ValueError(
-                f"Feedback for run_id={run_id!r} was already recorded as "
-                f"{conflicting_feedback.get('feedback_type')!r}"
-            )
-        if task.status in {"closed", "abandoned"} and not matching_feedback:
-            raise ValueError(f"Task {task.task_id} is already finalized as {task.status!r}")
-        if matching_feedback:
-            return task
-
-        entry = {
-            "feedback_type": feedback_type,
-            "comment": comment or "",
-            "run_id": run_id,
-            "created_at": now,
-        }
-        task.feedback.append(entry)
-        if feedback_type == "revise":
-            task.status = "needs_revision"
-        elif feedback_type == "abandon":
-            task.status = "abandoned"
-            task.closed_at = now
-            task.close_reason = comment or "abandoned"
-        elif feedback_type == "satisfied":
-            task.status = "closed"
-            task.closed_at = now
-            task.close_reason = "satisfied"
-            task.satisfaction = 1.0
-        task.updated_at = now
-        self.store.upsert_task(task)
-        self._event(task, f"feedback_{feedback_type}", run_id=run_id, payload=entry)
-        return task
 
     def close_task(self, task_id: str, *, reason: str = "closed") -> TaskRecord:
         task = self._require(task_id)
@@ -267,3 +259,12 @@ def short_task_title(text: str) -> str:
     if len(words) <= 4:
         return cleaned[:40]
     return " ".join(words[:4])[:40]
+
+
+def normalize_acceptance_type(value: str) -> str:
+    normalized = (value or "").strip().lower()
+    if normalized == "satisfied":
+        return "accept"
+    if normalized not in {"accept", "revise", "abandon"}:
+        raise ValueError("acceptance_type must be one of: accept, revise, abandon")
+    return normalized