fix: delete agent 500 error + dynamic personality + deployment guide

- Fix delete agent 500: clean up FK records (agent_llm_logs, permissions,
  schedules, executions, team_members) and unbind goals/tasks before delete
- Remove hardcoded personality templates in Android, replace with dynamic
  system prompt generation from name + description
- Set promptSectionsEnabled=false to bypass PromptComposer for personality
- Add Tencent Cloud Linux deployment guide (Docker Compose)
- Accumulated backend service updates, frontend UI fixes, Android app changes

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/app/services/workflow_engine.py

@@ -5578,21 +5578,14 @@ class WorkflowEngine:
                         "评分规则：1.0完美 0.8良好 0.6基本满足 0.4大部分未满足 0.2完全不满足\n"
                         "score >= 0.6 时 passed=true，否则 passed=false\n"
                     )
-                    messages = [{"role": "user", "content": judge_prompt}]
-                    resp = llm_service.call_openai_with_tools(
-                        provider="deepseek",
+                    judge_text = await llm_service.call_openai_with_tools(
+                        prompt=judge_prompt,
+                        tools=[],
                         model="deepseek-v4-flash",
-                        messages=messages,
                         temperature=0.1,
                         max_tokens=600,
                     )
-                    judge_text = ""
-                    if isinstance(resp, dict):
-                        choices = resp.get("choices", [])
-                        if choices:
-                            judge_text = choices[0].get("message", {}).get("content", "")
-                    elif hasattr(resp, 'choices'):
-                        judge_text = resp.choices[0].message.content or ""
+                    judge_text = str(judge_text or "")
 
                     # 解析 JSON
                     import re as _eval_re
@@ -5622,13 +5615,24 @@ class WorkflowEngine:
                             'summary': review.get('summary', ''),
                             'content': content if isinstance(content, str) else str(content),
                         },
-                        'status': 'success' if passed else 'quality_failed',
+                        'status': 'success',
+                        'branch': 'true' if passed else 'false',
                     }
                 except Exception as e:
                     logger.error(f"evaluator 节点评审失败: {e}")
                     exec_result = {
-                        'output': {'score': 0, 'passed': False, 'error': str(e), 'content': str(content)},
-                        'status': 'failed',
+                        'output': {
+                            'score': 0,
+                            'passed': False,
+                            'threshold': pass_threshold,
+                            'issues': [str(e)],
+                            'suggestions': [],
+                            'summary': f'评审执行异常: {e}',
+                            'error': str(e),
+                            'content': str(content),
+                        },
+                        'status': 'success',
+                        'branch': 'false',
                     }
 
                 if self.logger:
@@ -5843,8 +5847,8 @@ class WorkflowEngine:
                     if node.get('type') == 'start':
                         logger.info(f"[rjb] Start节点输出已保存: node_id={next_node_id}, output={output_value}, output_type={type(output_value)}")
                 
-                    # 如果是条件节点或Switch节点，根据分支结果过滤边
-                    if node.get('type') == 'condition':
+                    # 如果是条件节点/Switch节点/评估节点，根据分支结果过滤边
+                    if node.get('type') in ('condition', 'evaluator'):
                         branch = result.get('branch', 'false')
                         logger.info(f"[rjb] 条件节点分支过滤: node_id={next_node_id}, branch={branch}")
                         # 移除不符合条件的边