feat(task): route validation status to review states

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

@@ -110,10 +110,30 @@ class TaskService:
         self._event(task, "run_completed", run_id=run_id, payload={"skill_names": skill_names or []})
         return task
 
-    def record_validation(self, task_id: str, run_id: str, validation: ValidationResult) -> TaskRecord:
+    def record_validation(
+        self,
+        task_id: str,
+        run_id: str,
+        validation: ValidationResult,
+        *,
+        final_attempt: bool = True,
+        has_usable_answer: bool = True,
+    ) -> TaskRecord:
         task = self._require(task_id)
-        task.status = "awaiting_feedback"
-        task.updated_at = self._now()
+        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:
+            task.status = "needs_revision"
+        elif validation.status == "rejected" and has_usable_answer:
+            task.status = "needs_review"
+        else:
+            task.status = "failed"
+            task.closed_at = now
+            task.close_reason = "automatic validation rejected the final attempt"
+        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())

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

@@ -45,13 +45,13 @@ class ValidationService:
                 )
             except Exception as exc:
                 return ValidationResult(
-                    passed=False,
+                    status="validator_error",
                     score=0.0,
                     issues=[f"Validator failed: {exc}"],
-                    missing_requirements=["A valid automatic validation result is required before accepting the task."],
+                    evidence_gaps=["Automatic validation failed before producing a reliable decision."],
+                    missing_requirements=["User review is required because automatic validation failed."],
                     recommended_revision_prompt=(
-                        "Review the task result again because automatic validation failed, "
-                        "then provide a corrected final answer that explicitly satisfies the task goal."
+                        "Review the answer and evidence, then decide whether to revise or accept it."
                     ),
                     validator="llm_error",
                 )
@@ -96,11 +96,19 @@ class ValidationService:
             temperature=0.0,
         )
         payload = self._parse_json_object(response.content or "")
+        status = payload.get("status")
+        if status not in {"accepted", "rejected", "insufficient_evidence", "validator_error"}:
+            status = (
+                "accepted"
+                if payload.get("passed") and float(payload.get("score", 0.0) or 0.0) >= 0.75
+                else "rejected"
+            )
         return ValidationResult(
-            passed=bool(payload.get("passed")),
+            status=status,
             score=max(0.0, min(1.0, float(payload.get("score", 0.0) or 0.0))),
             issues=[str(item) for item in payload.get("issues") or []],
             missing_requirements=[str(item) for item in payload.get("missing_requirements") or []],
+            evidence_gaps=[str(item) for item in payload.get("evidence_gaps") or []],
             recommended_revision_prompt=str(payload.get("recommended_revision_prompt") or ""),
             validator="llm",
         )