openclaw-agent-dashboard 1.0.39 → 1.0.40

package/dashboard/api/input_safety.py

@@ -0,0 +1,60 @@
+"""
+外部路径与 Query 边界校验（NFR-S-002）。
+拒绝路径逃逸与过长输入，避免异常 agent_id / session 片段误触文件逻辑。
+"""
+from __future__ import annotations
+
+from fastapi import HTTPException
+
+_MAX_AGENT_ID_LEN = 128
+_MAX_SESSION_KEY_LEN = 512
+_MAX_RUN_CHAIN_ID_LEN = 128
+_MAX_SESSION_FILE_SEGMENT = 256
+
+
+def require_safe_agent_id(agent_id: str) -> str:
+    s = (agent_id or "").strip()
+    if not s or len(s) > _MAX_AGENT_ID_LEN:
+        raise HTTPException(status_code=400, detail="invalid agent_id")
+    if "\x00" in s:
+        raise HTTPException(status_code=400, detail="invalid agent_id")
+    if ".." in s:
+        raise HTTPException(status_code=400, detail="invalid agent_id")
+    for ch in ("/", "\\"):
+        if ch in s:
+            raise HTTPException(status_code=400, detail="invalid agent_id")
+    low = s.lower()
+    if "%2f" in low or "%5c" in low or "%2e%2e" in low:
+        raise HTTPException(status_code=400, detail="invalid agent_id")
+    return s
+
+
+def require_safe_session_key(session_key: str | None) -> str | None:
+    if session_key is None:
+        return None
+    s = session_key.strip()
+    if not s:
+        return None
+    if len(s) > _MAX_SESSION_KEY_LEN or "\x00" in s:
+        raise HTTPException(status_code=400, detail="invalid session_key")
+    if ".." in s or "/" in s or "\\" in s:
+        raise HTTPException(status_code=400, detail="invalid session_key")
+    return s
+
+
+def require_safe_session_file_segment(session_file: str) -> str:
+    s = (session_file or "").strip()
+    if not s or len(s) > _MAX_SESSION_FILE_SEGMENT:
+        raise HTTPException(status_code=400, detail="invalid session_file")
+    if "\x00" in s or ".." in s or "/" in s or "\\" in s:
+        raise HTTPException(status_code=400, detail="invalid session_file")
+    return s
+
+
+def require_safe_run_or_chain_id(value: str, *, name: str = "id") -> str:
+    s = (value or "").strip()
+    if not s or len(s) > _MAX_RUN_CHAIN_ID_LEN:
+        raise HTTPException(status_code=400, detail=f"invalid {name}")
+    if "\x00" in s or ".." in s or "/" in s or "\\" in s:
+        raise HTTPException(status_code=400, detail=f"invalid {name}")
+    return s

package/dashboard/api/performance.py

@@ -10,7 +10,9 @@ from pathlib import Path
 from datetime import datetime, timedelta, timezone
 from zoneinfo import ZoneInfo
 
-from data.session_reader import normalize_sessions_index
+from data.session_reader import normalize_sessions_index, _load_sessions_index_file
+from core.error_handler import record_error
+from utils.data_repair import parse_session_jsonl_line
 
 # 详情展示使用 Asia/Shanghai 时区
 TZ_DISPLAY = ZoneInfo('Asia/Shanghai')
@@ -103,23 +105,23 @@ def parse_session_file_with_details(session_path: Path, agent_id: str) -> List[D
         with open(session_path, 'r', encoding='utf-8') as f:
             for line in f:
                 try:
-                    data = json.loads(line)
-                    if data.get('type') != 'message':
+                    data, msg = parse_session_jsonl_line(line)
+                    if not data or data.get('type') != 'message' or not msg:
                         continue
-                    msg = data.get('message', {})
-                    if not msg:
-                        continue
-                    
+
                     msg_id = data.get('id', '')
                     id_to_msg[msg_id] = {'data': data, 'msg': msg}
-                    
+
                     if msg.get('role') != 'assistant':
                         continue
                     if 'usage' not in msg:
                         continue
-                    
+
+                    ts_raw = data.get('timestamp')
+                    if not ts_raw:
+                        continue
                     try:
-                        ts = datetime.fromisoformat(data['timestamp'].replace('Z', '+00:00'))
+                        ts = datetime.fromisoformat(str(ts_raw).replace('Z', '+00:00'))
                     except Exception:
                         continue
                     
@@ -163,7 +165,7 @@ def parse_session_file_with_details(session_path: Path, agent_id: str) -> List[D
                     continue
         return records
     except Exception as e:
-        print(f"解析 session 详情失败 {session_path}: {e}")
+        record_error("io-error", f"{session_path}: {e}", "performance:parse_session_details", exc=e)
         return []
 
 
@@ -180,37 +182,43 @@ def parse_session_file(session_path: Path, range_hours: int = 1) -> List[Dict]:
         with open(session_path, 'r', encoding='utf-8') as f:
             for line in f:
                 try:
-                    data = json.loads(line)
-
-                    # 只处理有 usage 和 timestamp 的消息
-                    if 'message' in data and 'usage' in data['message'] and 'timestamp' in data:
-                        usage = data['message']['usage']
-                        tokens = usage.get('totalTokens', 0) or 0
-                        is_request = data.get('message', {}).get('role') == 'assistant'
-
-                        try:
-                            timestamp = datetime.fromisoformat(data['timestamp'].replace('Z', '+00:00'))
-
-                            # 根据 range_hours 过滤时间范围，0 表示不过滤
-                            if range_hours > 0:
-                                now = datetime.now(timezone.utc)
-                                time_ago = now - timedelta(hours=range_hours)
-                                if timestamp < time_ago:
-                                    continue
+                    envelope, msg = parse_session_jsonl_line(line)
+                    if (
+                        not envelope
+                        or envelope.get('type') != 'message'
+                        or not msg
+                        or 'usage' not in msg
+                        or not envelope.get('timestamp')
+                    ):
+                        continue
+                    usage = msg['usage']
+                    tokens = usage.get('totalTokens', 0) or 0
+                    is_request = msg.get('role') == 'assistant'
 
-                            messages.append({
-                                'timestamp': timestamp,
-                                'tokens': tokens,
-                                'is_request': is_request
-                            })
-                        except:
-                            pass
-                except:
+                    try:
+                        timestamp = datetime.fromisoformat(
+                            str(envelope['timestamp']).replace('Z', '+00:00')
+                        )
+
+                        if range_hours > 0:
+                            now = datetime.now(timezone.utc)
+                            time_ago = now - timedelta(hours=range_hours)
+                            if timestamp < time_ago:
+                                continue
+
+                        messages.append({
+                            'timestamp': timestamp,
+                            'tokens': tokens,
+                            'is_request': is_request
+                        })
+                    except Exception:
+                        pass
+                except Exception:
                     continue
 
         return messages
     except Exception as e:
-        print(f"解析 session 文件失败 {session_path}: {e}")
+        record_error("io-error", f"{session_path}: {e}", "performance:parse_session_file", exc=e)
         return []
 
 
@@ -491,9 +499,7 @@ async def get_minute_details(
             }
         }
     except Exception as e:
-        print(f"获取调用详情失败: {e}")
-        import traceback
-        traceback.print_exc()
+        record_error("unknown", str(e), "performance:get_minute_details", exc=e)
         return {'timeWindow': '', 'calls': [], 'totalCalls': 0, 'totalTokens': 0, 'summary': {'avgTokens': 0}, 'agents': [], 'pagination': {'total': 0, 'limit': limit, 'hasMore': False}}
 
 
@@ -629,19 +635,33 @@ async def get_tokens_analysis(range: str = "all"):
                     with open(session_file, 'r', encoding='utf-8') as f:
                         for line in f:
                             try:
-                                data = json.loads(line)
-                                if data.get('type') != 'message':
+                                envelope, msg = parse_session_jsonl_line(line)
+                                if (
+                                    envelope is None
+                                    or envelope.get('type') != 'message'
+                                    or msg is None
+                                ):
                                     continue
-                                msg = data.get('message', {})
                                 if msg.get('role') != 'assistant' or 'usage' not in msg:
                                     continue
 
-                                try:
-                                    ts = datetime.fromisoformat(data['timestamp'].replace('Z', '+00:00'))
-                                except:
-                                    continue
-
-                                if ts < time_ago:
+                                ts_raw = envelope.get('timestamp') or msg.get('timestamp')
+                                ts = None
+                                if isinstance(ts_raw, (int, float)):
+                                    v = float(ts_raw)
+                                    ts = datetime.fromtimestamp(
+                                        (v / 1000.0) if v > 1e12 else v, tz=timezone.utc
+                                    )
+                                elif isinstance(ts_raw, str):
+                                    try:
+                                        ts = datetime.fromisoformat(
+                                            ts_raw.replace('Z', '+00:00')
+                                        )
+                                        if ts.tzinfo is None:
+                                            ts = ts.replace(tzinfo=timezone.utc)
+                                    except ValueError:
+                                        ts = None
+                                if ts is None or ts < time_ago:
                                     continue
 
                                 usage = msg['usage']
@@ -706,9 +726,8 @@ async def get_tokens_analysis(range: str = "all"):
                 continue
 
             try:
-                with open(sessions_index, 'r', encoding='utf-8') as f:
-                    data = json.load(f)
-                if not isinstance(data, dict):
+                data = _load_sessions_index_file(sessions_index)
+                if not data:
                     continue
 
                 agent_total = {"input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0}
@@ -739,7 +758,8 @@ async def get_tokens_analysis(range: str = "all"):
                 result["summary"]["output"] += agent_total["output"]
                 result["summary"]["cacheRead"] += agent_total["cacheRead"]
                 result["summary"]["cacheWrite"] += agent_total["cacheWrite"]
-            except Exception:
+            except Exception as e:
+                record_error("unknown", str(e), "performance:tokens_analysis_agent", exc=e)
                 continue
 
     # 计算汇总

package/dashboard/api/subagents.py

@@ -17,7 +17,14 @@ from data.subagent_reader import (
     get_agent_files_for_run
 )
 from data.task_history import merge_with_history
-from data.session_reader import normalize_sessions_index, resolve_session_jsonl_path
+from data.session_reader import (
+    normalize_sessions_index,
+    resolve_session_jsonl_path,
+    _load_sessions_index_file,
+)
+from utils.data_repair import parse_session_jsonl_line
+from core.error_handler import record_error
+from core.safe_api_error import safe_client_string
 import time
 
 router = APIRouter()
@@ -108,9 +115,7 @@ async def get_subagents():
 
         return result
     except Exception as e:
-        print(f"Error in get_subagents: {e}")
-        import traceback
-        traceback.print_exc()
+        record_error("unknown", str(e), "api:subagents:get_subagents", exc=e)
         return []
 
 
@@ -139,7 +144,7 @@ async def get_active_subagents():
 
         return result
     except Exception as e:
-        print(f"Error in get_active_subagents: {e}")
+        record_error("unknown", str(e), "api:subagents:get_active_subagents", exc=e)
         return []
 
 
@@ -149,7 +154,8 @@ def _get_agent_name(agent_id: str) -> str:
         from data.config_reader import get_agent_config
         config = get_agent_config(agent_id)
         return config.get('name', agent_id) if config else agent_id
-    except Exception:
+    except Exception as e:
+        record_error("unknown", str(e), "api:subagents:agent_name", exc=e)
         return agent_id
 
 
@@ -161,7 +167,8 @@ def _get_agent_workspace(agent_id: str) -> Optional[str]:
         from data.config_reader import get_agent_config
         config = get_agent_config(agent_id)
         return config.get('workspace') if config else None
-    except Exception:
+    except Exception as e:
+        record_error("unknown", str(e), "api:subagents:workspace", exc=e)
         return None
 
 
@@ -269,8 +276,9 @@ def _get_session_message_count(child_session_key: str) -> int:
         if not sessions_index.exists():
             return 0
 
-        with open(sessions_index, 'r', encoding='utf-8') as f:
-            index_data = json.load(f)
+        index_data = _load_sessions_index_file(sessions_index)
+        if not index_data:
+            return 0
         index_map = normalize_sessions_index(index_data)
         entry = index_map.get(child_session_key)
         if not entry:
@@ -283,15 +291,12 @@ def _get_session_message_count(child_session_key: str) -> int:
         count = 0
         with open(session_path, 'r', encoding='utf-8') as f:
             for line in f:
-                try:
-                    data = json.loads(line)
-                    if data.get('type') == 'message':
-                        count += 1
-                except json.JSONDecodeError:
-                    continue
+                envelope, msg = parse_session_jsonl_line(line)
+                if envelope and envelope.get('type') == 'message' and msg is not None:
+                    count += 1
         return count
     except Exception as e:
-        print(f"_get_session_message_count 失败: {e}")
+        record_error("io-error", str(e), "api:subagents:session_message_count", exc=e)
         return 0
 
 
@@ -360,8 +365,9 @@ def _extract_subtasks_from_session(child_session_key: str) -> List[Dict[str, Any
         if not sessions_index.exists():
             return []
 
-        with open(sessions_index, 'r', encoding='utf-8') as f:
-            index_data = json.load(f)
+        index_data = _load_sessions_index_file(sessions_index)
+        if not index_data:
+            return []
         index_map = normalize_sessions_index(index_data)
         entry = index_map.get(child_session_key)
         if not entry:
@@ -377,10 +383,9 @@ def _extract_subtasks_from_session(child_session_key: str) -> List[Dict[str, Any
         with open(session_path, 'r', encoding='utf-8') as f:
             for line in f:
                 try:
-                    data = json.loads(line)
-                    if data.get('type') != 'message':
+                    envelope, msg = parse_session_jsonl_line(line)
+                    if not envelope or envelope.get('type') != 'message' or not msg:
                         continue
-                    msg = data.get('message', {})
                     if msg.get('role') != 'assistant':
                         continue
                     content = msg.get('content', [])
@@ -396,7 +401,6 @@ def _extract_subtasks_from_session(child_session_key: str) -> List[Dict[str, Any
                                 args = json.loads(args)
                             except json.JSONDecodeError:
                                 continue
-                        # 提取子任务信息
                         task_desc = args.get('task') or args.get('prompt') or args.get('instruction', '')
                         sub_agent_id = args.get('agentId') or args.get('agent') or args.get('agent_id', '')
                         if task_desc and task_desc not in seen_tasks:
@@ -404,14 +408,14 @@ def _extract_subtasks_from_session(child_session_key: str) -> List[Dict[str, Any
                             subtasks.append({
                                 'task': task_desc[:200] if len(task_desc) > 200 else task_desc,
                                 'agentId': sub_agent_id,
-                                'status': 'unknown'  # 无法从 session 确定状态
+                                'status': 'unknown'
                             })
-                except (json.JSONDecodeError, KeyError):
+                except (KeyError, TypeError):
                     continue
 
         return subtasks[:5]  # 最多返回 5 个子任务
     except Exception as e:
-        print(f"_extract_subtasks_from_session 失败: {e}")
+        record_error("io-error", str(e), "api:subagents:extract_subtasks", exc=e)
         return []
 
 
@@ -490,9 +494,7 @@ async def get_tasks():
                 t['agentWorkspace'] = _get_agent_workspace(t['agentId'])
         return {'tasks': tasks}
     except Exception as e:
-        print(f"Error in get_tasks: {e}")
-        import traceback
-        traceback.print_exc()
+        record_error("unknown", str(e), "api:subagents:get_tasks", exc=e)
         return {'tasks': []}
 
 
@@ -521,8 +523,9 @@ def _extract_timeline_from_session(child_session_key: str) -> List[Dict[str, Any
         if not sessions_index.exists():
             return []
 
-        with open(sessions_index, 'r', encoding='utf-8') as f:
-            index_data = json.load(f)
+        index_data = _load_sessions_index_file(sessions_index)
+        if not index_data:
+            return []
         index_map = normalize_sessions_index(index_data)
         entry = index_map.get(child_session_key)
         if not entry:
@@ -541,16 +544,15 @@ def _extract_timeline_from_session(child_session_key: str) -> List[Dict[str, Any
                 if event_count >= max_events:
                     break
                 try:
-                    data = json.loads(line)
-                    ts = data.get('timestamp')
-                    # 确保 ts 是整数毫秒时间戳
+                    envelope, msg = parse_session_jsonl_line(line)
+                    if not envelope or envelope.get('type') != 'message' or not msg:
+                        continue
+                    ts = envelope.get('timestamp')
                     if isinstance(ts, str):
-                        # ISO 格式转毫秒时间戳
                         try:
-                            from datetime import datetime
-                            # 处理 ISO 格式：2026-03-07T04:07:25.262Z
+                            from datetime import datetime as _dt
                             if 'T' in ts:
-                                dt = datetime.fromisoformat(ts.replace('Z', '+00:00'))
+                                dt = _dt.fromisoformat(ts.replace('Z', '+00:00'))
                                 ts = int(dt.timestamp() * 1000)
                             else:
                                 ts = int(ts)
@@ -561,68 +563,61 @@ def _extract_timeline_from_session(child_session_key: str) -> List[Dict[str, Any
                     else:
                         ts = 0
 
-                    if data.get('type') == 'message':
-                        msg = data.get('message', {})
-                        role = msg.get('role', '')
-                        content = msg.get('content', [])
-
-                        if role == 'user':
-                            # 用户消息（任务开始）
-                            for c in content:
-                                if isinstance(c, dict) and c.get('type') == 'text':
-                                    text = c.get('text', '')[:100]
-                                    timeline.append({
-                                        'time': ts,
-                                        'type': 'start',
-                                        'description': f'收到任务: {text}...' if len(c.get('text', '')) > 100 else f'收到任务: {text}'
-                                    })
-                                    event_count += 1
-                                    break
-
-                        elif role == 'assistant':
-                            # 助手响应中的工具调用
-                            for c in content:
-                                if not isinstance(c, dict):
-                                    continue
-                                if c.get('type') == 'toolCall':
-                                    tool_name = c.get('name', 'unknown')
-                                    args = c.get('arguments', {})
-                                    if isinstance(args, str):
-                                        try:
-                                            args = json.loads(args)
-                                        except json.JSONDecodeError:
-                                            args = {}
-
-                                    # 生成描述
-                                    desc = _describe_tool_call(tool_name, args)
-                                    timeline.append({
-                                        'time': ts,
-                                        'type': 'tool',
-                                        'tool': tool_name,
-                                        'description': desc
-                                    })
-                                    event_count += 1
-
-                                elif c.get('type') == 'text':
-                                    # 文本响应（可能是最终答案）
-                                    text = c.get('text', '')
-                                    if text.strip() and len(text) > 50:
-                                        # 简单判断是否是最终答案
-                                        keywords = ['完成', '成功', 'finished', 'done', 'result', '总结']
-                                        if any(kw in text.lower() for kw in keywords):
-                                            timeline.append({
-                                                'time': ts,
-                                                'type': 'response',
-                                                'description': f'输出结果 ({len(text)} 字符)'
-                                            })
-                                            event_count += 1
-
-                except (json.JSONDecodeError, KeyError):
+                    role = msg.get('role', '')
+                    content = msg.get('content', [])
+
+                    if role == 'user':
+                        for c in content:
+                            if isinstance(c, dict) and c.get('type') == 'text':
+                                text = c.get('text', '')[:100]
+                                timeline.append({
+                                    'time': ts,
+                                    'type': 'start',
+                                    'description': f'收到任务: {text}...' if len(c.get('text', '')) > 100 else f'收到任务: {text}'
+                                })
+                                event_count += 1
+                                break
+
+                    elif role == 'assistant':
+                        for c in content:
+                            if not isinstance(c, dict):
+                                continue
+                            if c.get('type') == 'toolCall':
+                                tool_name = c.get('name', 'unknown')
+                                args = c.get('arguments', {})
+                                if isinstance(args, str):
+                                    try:
+                                        args = json.loads(args)
+                                    except json.JSONDecodeError:
+                                        args = {}
+
+                                desc = _describe_tool_call(tool_name, args)
+                                timeline.append({
+                                    'time': ts,
+                                    'type': 'tool',
+                                    'tool': tool_name,
+                                    'description': desc
+                                })
+                                event_count += 1
+
+                            elif c.get('type') == 'text':
+                                text = c.get('text', '')
+                                if text.strip() and len(text) > 50:
+                                    keywords = ['完成', '成功', 'finished', 'done', 'result', '总结']
+                                    if any(kw in text.lower() for kw in keywords):
+                                        timeline.append({
+                                            'time': ts,
+                                            'type': 'response',
+                                            'description': f'输出结果 ({len(text)} 字符)'
+                                        })
+                                        event_count += 1
+
+                except (KeyError, TypeError, ValueError):
                     continue
 
         return timeline
     except Exception as e:
-        print(f"_extract_timeline_from_session 失败: {e}")
+        record_error("io-error", str(e), "api:subagents:extract_timeline", exc=e)
         return []
 
 
@@ -674,6 +669,9 @@ async def get_task_timeline(run_id: str):
     Returns:
         时间线事件列表
     """
+    from api.input_safety import require_safe_run_or_chain_id
+
+    require_safe_run_or_chain_id(run_id, name="run_id")
     try:
         # 从 runs.json 查找对应的 session key
         all_runs = load_subagent_runs()
@@ -750,7 +748,5 @@ async def get_task_timeline(run_id: str):
 
         return {'timeline': timeline, 'runId': run_id}
     except Exception as e:
-        print(f"Error in get_task_timeline: {e}")
-        import traceback
-        traceback.print_exc()
-        return {'timeline': [], 'error': str(e)}
+        record_error("unknown", str(e), "api:subagents:get_task_timeline", exc=e)
+        return {'timeline': [], 'error': safe_client_string(str(e))}

package/dashboard/api/timeline.py

@@ -12,6 +12,9 @@ from pathlib import Path
 LOG = logging.getLogger(__name__)
 sys.path.append(str(Path(__file__).parent.parent))
 
+from api.input_safety import require_safe_agent_id, require_safe_session_key
+from core.error_handler import record_error
+from core.safe_api_error import safe_api_error_detail
 from data.timeline_reader import get_timeline_steps, StepType, StepStatus
 from data.config_reader import get_agent_config
 
@@ -70,12 +73,18 @@ async def get_timeline(
     - 工具调用及结果
     - 错误信息
     """
+    require_safe_agent_id(agent_id)
+    session_key = require_safe_session_key(session_key)
     agent_info = get_agent_config(agent_id)
     if not agent_info:
         raise HTTPException(status_code=404, detail=f"Agent {agent_id} not found")
 
     t0 = time.perf_counter()
-    result = get_timeline_steps(agent_id, session_key, limit)
+    try:
+        result = get_timeline_steps(agent_id, session_key, limit)
+    except Exception as e:
+        record_error("unknown", str(e), "api:timeline:get", exc=e)
+        raise HTTPException(status_code=500, detail=safe_api_error_detail(e)) from e
     elapsed_ms = (time.perf_counter() - t0) * 1000
     if elapsed_ms >= 200.0:
         LOG.info(
@@ -111,10 +120,16 @@ async def get_timeline_steps_only(
 
     可按步骤类型过滤
     """
+    require_safe_agent_id(agent_id)
+    session_key = require_safe_session_key(session_key)
     if not get_agent_config(agent_id):
         raise HTTPException(status_code=404, detail=f"Agent {agent_id} not found")
 
-    result = get_timeline_steps(agent_id, session_key, limit)
+    try:
+        result = get_timeline_steps(agent_id, session_key, limit)
+    except Exception as e:
+        record_error("unknown", str(e), "api:timeline:steps", exc=e)
+        raise HTTPException(status_code=500, detail=safe_api_error_detail(e)) from e
     steps = result.get('steps', [])
 
     # 类型过滤
@@ -131,10 +146,16 @@ async def get_timeline_summary(agent_id: str, session_key: Optional[str] = Query
 
     快速查看会话概览，不返回详细步骤
     """
+    require_safe_agent_id(agent_id)
+    session_key = require_safe_session_key(session_key)
     if not get_agent_config(agent_id):
         raise HTTPException(status_code=404, detail=f"Agent {agent_id} not found")
 
-    result = get_timeline_steps(agent_id, session_key, limit=10)  # 只需基本信息
+    try:
+        result = get_timeline_steps(agent_id, session_key, limit=10)  # 只需基本信息
+    except Exception as e:
+        record_error("unknown", str(e), "api:timeline:summary", exc=e)
+        raise HTTPException(status_code=500, detail=safe_api_error_detail(e)) from e
 
     # 统计各类型步骤数量
     steps = result.get('steps', [])