openclaw-agent-dashboard 1.0.39 → 1.0.40

package/dashboard/data/subagent_reader.py

@@ -6,7 +6,16 @@ from pathlib import Path
 from typing import List, Dict, Any, Optional
 
 from data.config_reader import get_openclaw_root, normalize_openclaw_agent_id
-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 core.config_fortify import get_fortify_config
+from core.error_handler import record_error
+from core.schemas.base import SchemaValidator
+from core.schemas.subagent_schema import subagent_runs_root_schema
+from utils.data_repair import parse_session_jsonl_line
 
 
 def load_subagent_runs() -> List[Dict[str, Any]]:
@@ -17,13 +26,35 @@ def load_subagent_runs() -> List[Dict[str, Any]]:
     runs_path = get_openclaw_root() / "subagents" / "runs.json"
     if not runs_path.exists():
         return []
-    
-    with open(runs_path, 'r', encoding='utf-8') as f:
-        data = json.load(f)
-    
+
+    try:
+        with open(runs_path, 'r', encoding='utf-8') as f:
+            data = json.load(f)
+    except (json.JSONDecodeError, OSError) as e:
+        record_error("parsing-error", str(e), "subagent_runs")
+        return []
+
+    if not isinstance(data, dict):
+        return []
+
+    cfg = get_fortify_config()
+    vr = SchemaValidator(subagent_runs_root_schema, strict=cfg.json_strict).validate(data)
+    if not vr.is_valid:
+        record_error("validation-error", vr.error_message, "subagent_runs")
+        if cfg.json_strict:
+            return []
+
     runs = data.get('runs', {})
     if isinstance(runs, dict):
-        return list(runs.values())
+        out: List[Dict[str, Any]] = []
+        for run_id, rec in runs.items():
+            if not isinstance(rec, dict):
+                continue
+            merged = dict(rec)
+            if not merged.get('runId'):
+                merged['runId'] = run_id
+            out.append(merged)
+        return out
     return runs if isinstance(runs, list) else []
 
 
@@ -118,8 +149,9 @@ def get_agent_output_for_run(child_session_key: str, max_chars: int = 10000) ->
         return None
     
     try:
-        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 None
         index_map = normalize_sessions_index(index_data)
         entry = index_map.get(child_session_key)
         if not entry:
@@ -132,22 +164,18 @@ def get_agent_output_for_run(child_session_key: str, max_chars: int = 10000) ->
         last_text = None
         with open(session_path, 'r', encoding='utf-8') as f:
             for line in f:
-                try:
-                    data = json.loads(line)
-                    if data.get('type') != 'message':
-                        continue
-                    msg = data.get('message', {})
-                    if msg.get('role') != 'assistant':
-                        continue
-                    content = msg.get('content', [])
-                    for c in content:
-                        if isinstance(c, dict) and c.get('type') == 'text':
-                            text = c.get('text', '')
-                            if text and text.strip():
-                                last_text = text
-                            break
-                except (json.JSONDecodeError, KeyError):
+                envelope, msg = parse_session_jsonl_line(line)
+                if not envelope or envelope.get('type') != 'message' or msg is None:
                     continue
+                if msg.get('role') != 'assistant':
+                    continue
+                content = msg.get('content', [])
+                for c in content:
+                    if isinstance(c, dict) and c.get('type') == 'text':
+                        text = c.get('text', '')
+                        if text and text.strip():
+                            last_text = text
+                        break
         
         if not last_text or not last_text.strip():
             return None
@@ -155,7 +183,7 @@ def get_agent_output_for_run(child_session_key: str, max_chars: int = 10000) ->
             return last_text[:max_chars] + '\n\n...(输出已截断)'
         return last_text
     except Exception as e:
-        print(f"get_agent_output_for_run 失败: {e}")
+        record_error("io-error", str(e), "subagent_reader:get_agent_output_for_run", exc=e)
         return None
 
 
@@ -180,8 +208,9 @@ def get_agent_files_for_run(child_session_key: str) -> List[str]:
         return []
     
     try:
-        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:
@@ -196,31 +225,27 @@ def get_agent_files_for_run(child_session_key: str) -> List[str]:
         
         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 msg is None:
+                    continue
+                if msg.get('role') != 'assistant':
+                    continue
+                content = msg.get('content', [])
+                for c in content:
+                    if not isinstance(c, dict) or c.get('type') != 'toolCall':
                         continue
-                    msg = data.get('message', {})
-                    if msg.get('role') != 'assistant':
+                    name = c.get('name', '')
+                    if name not in file_tools:
                         continue
-                    content = msg.get('content', [])
-                    for c in content:
-                        if not isinstance(c, dict) or c.get('type') != 'toolCall':
+                    args = c.get('arguments', {})
+                    if isinstance(args, str):
+                        try:
+                            args = json.loads(args)
+                        except json.JSONDecodeError:
                             continue
-                        name = c.get('name', '')
-                        if name not in file_tools:
-                            continue
-                        args = c.get('arguments', {})
-                        if isinstance(args, str):
-                            try:
-                                args = json.loads(args)
-                            except json.JSONDecodeError:
-                                continue
-                        path = args.get('path') or args.get('file_path')
-                        if path and isinstance(path, str) and path.strip():
-                            file_paths.append(path.strip())
-                except (json.JSONDecodeError, KeyError):
-                    continue
+                    path = args.get('path') or args.get('file_path')
+                    if path and isinstance(path, str) and path.strip():
+                        file_paths.append(path.strip())
         
         # 去重并保持顺序
         seen = set()
@@ -231,5 +256,5 @@ def get_agent_files_for_run(child_session_key: str) -> List[str]:
                 result.append(p)
         return result
     except Exception as e:
-        print(f"get_agent_files_for_run 失败: {e}")
+        record_error("io-error", str(e), "subagent_reader:get_agent_files_for_run", exc=e)
         return []

package/dashboard/data/timeline_reader.py

@@ -1,7 +1,6 @@
 """
 时序数据读取器 - 将 session jsonl 解析为可视化时序步骤
 """
-import json
 import logging
 import os
 from functools import lru_cache
@@ -92,7 +91,13 @@ class TimelineStep:
 
 
 from data.config_reader import get_openclaw_root, normalize_openclaw_agent_id, agent_ids_equal
-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 data.subagent_reader import load_subagent_runs
+from utils.data_repair import parse_session_jsonl_line
 
 
 def _read_session_header_timestamp(path: Path) -> Optional[int]:
@@ -102,10 +107,10 @@ def _read_session_header_timestamp(path: Path) -> Optional[int]:
             first = f.readline()
         if not first.strip():
             return None
-        data = json.loads(first.strip())
-        if data.get('type') == 'session':
-            return _parse_timestamp(data.get('timestamp', 0))
-    except (json.JSONDecodeError, OSError, IOError):
+        envelope, _ = parse_session_jsonl_line(first.strip())
+        if envelope and envelope.get('type') == 'session':
+            return _parse_timestamp(envelope.get('timestamp', 0))
+    except (OSError, IOError):
         pass
     return None
 
@@ -260,17 +265,11 @@ def get_subagent_runs() -> Dict[str, List[Dict]]:
 
 @lru_cache(maxsize=16)
 def _get_subagent_runs_cached(mtime: float) -> Dict[str, List[Dict]]:
-    runs_file = get_openclaw_root() / "subagents" / "runs.json"
-    try:
-        with open(runs_file, 'r', encoding='utf-8') as f:
-            data = json.load(f)
-    except (json.JSONDecodeError, IOError, OSError):
-        return {}
     runs_by_agent: Dict[str, List[Dict]] = {}
-    runs = data.get('runs', {})
-    for run_id, run_info in runs.items():
+    for run_info in load_subagent_runs():
         if not isinstance(run_info, dict):
             continue
+        run_id = str(run_info.get('runId', ''))
         child_key = run_info.get('childSessionKey', '')
         if ':' in child_key:
             parts = child_key.split(':')
@@ -297,9 +296,11 @@ def _get_requester_info_for_session(agent_id: str, session_key: Optional[str]) -
     sessions_index = get_openclaw_root() / f"agents/{state_id}/sessions/sessions.json"
     if sessions_index.exists():
         try:
-            with open(sessions_index, 'r', encoding='utf-8') as f:
-                index_data = json.load(f)
-            index_map = normalize_sessions_index(index_data)
+            index_data = _load_sessions_index_file(sessions_index)
+            if not index_data:
+                index_map = {}
+            else:
+                index_map = normalize_sessions_index(index_data)
             if not session_key:
                 entries = list(index_map.items())
                 if entries:
@@ -328,14 +329,8 @@ def _get_requester_info_for_session(agent_id: str, session_key: Optional[str]) -
             session_key = runs[0].get('childSessionKey')
     if not session_key:
         return {}
-    runs_file = get_openclaw_root() / "subagents" / "runs.json"
-    if not runs_file.exists():
-        return {}
     try:
-        with open(runs_file, 'r', encoding='utf-8') as f:
-            data = json.load(f)
-        runs = data.get('runs', {})
-        for run_id, run_info in runs.items():
+        for run_info in load_subagent_runs():
             if not isinstance(run_info, dict):
                 continue
             child_key = run_info.get('childSessionKey', '')
@@ -612,10 +607,10 @@ def resolve_agent_session_jsonl(
     index_path = sessions_path / "sessions.json"
     index_map: Dict[str, Dict[str, Any]] = {}
     if index_path.exists():
-        try:
-            with open(index_path, 'r', encoding='utf-8') as f:
-                index_map = normalize_sessions_index(json.load(f))
-        except (json.JSONDecodeError, IOError):
+        raw = _load_sessions_index_file(index_path)
+        if raw:
+            index_map = normalize_sessions_index(raw)
+        else:
             index_map = {}
 
     prefix = f"agent:{state_id}:"
@@ -895,17 +890,13 @@ def _extract_subagent_steps_from_main_lines(
     for line in lines:
         if '"type":"message"' not in line and '"type": "message"' not in line:
             continue
-        try:
-            data = json.loads(line.strip())
-        except json.JSONDecodeError:
-            continue
-        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', {})
         role = msg.get('role')
         if not role:
             continue
-        timestamp = _parse_timestamp(msg.get('timestamp') or data.get('timestamp', 0))
+        timestamp = _parse_timestamp(msg.get('timestamp') or envelope.get('timestamp', 0))
         duration = 0
         if last_timestamp and timestamp:
             duration = timestamp - last_timestamp
@@ -1112,21 +1103,19 @@ def _parse_session_lines(
     sender_id = requester_info.get('senderId') if requester_info else None
     sender_name = requester_info.get('senderName') if requester_info else None
     for line in lines:
-        try:
-            data = json.loads(line.strip())
-        except json.JSONDecodeError:
+        envelope, msg = parse_session_jsonl_line(line)
+        if envelope is None:
             continue
-        msg_type = data.get('type')
+        msg_type = envelope.get('type')
         if msg_type == 'session':
-            started_at = _parse_timestamp(data.get('timestamp', 0))
+            started_at = _parse_timestamp(envelope.get('timestamp', 0))
             continue
-        if msg_type != 'message':
+        if msg_type != 'message' or msg is None:
             continue
-        msg = data.get('message', {})
         role = msg.get('role')
         if not role:
             continue
-        timestamp = _parse_timestamp(msg.get('timestamp') or data.get('timestamp', 0))
+        timestamp = _parse_timestamp(msg.get('timestamp') or envelope.get('timestamp', 0))
         duration = 0
         if last_timestamp and timestamp:
             duration = timestamp - last_timestamp
@@ -1295,13 +1284,10 @@ def _line_index_of_first_user_message(path: Path) -> Optional[int]:
             for i, line in enumerate(f):
                 if '"role"' not in line or 'user' not in line:
                     continue
-                try:
-                    d = json.loads(line.strip())
-                except json.JSONDecodeError:
-                    continue
-                if d.get('type') != 'message':
+                env, msg = parse_session_jsonl_line(line)
+                if env is None or env.get('type') != 'message' or msg is None:
                     continue
-                if (d.get('message') or {}).get('role') == 'user':
+                if msg.get('role') == 'user':
                     return i
     except (OSError, IOError):
         pass

package/dashboard/main.py

@@ -13,14 +13,35 @@ import asyncio
 async def lifespan(app: FastAPI):
     """应用生命周期：启动时启动文件监听，关闭时停止"""
     loop = asyncio.get_running_loop()
+    probe_stop = None
     try:
         from watchers.file_watcher import start_file_watcher
+        from core.config_fortify import get_fortify_config
+
         start_file_watcher(loop)
+        cfg = get_fortify_config()
+        if cfg.cache_preload:
+            try:
+                from status.status_calculator import get_agents_with_status
+
+                get_agents_with_status()
+            except Exception as e:
+                from core.error_handler import record_error
+
+                record_error("unknown", str(e), "main:cache_preload", exc=e)
+        from status.cache_fp_probe import start_cache_fp_probe_background
+
+        probe_stop = start_cache_fp_probe_background()
     except Exception as e:
-        print(f"[Main] 文件监听启动失败: {e}")
+        from core.error_handler import record_error
+
+        record_error("unknown", str(e), "main:file_watcher_start", exc=e)
     yield
     try:
+        if probe_stop is not None:
+            probe_stop.set()
         from watchers.file_watcher import stop_file_watcher
+
         stop_file_watcher()
     except Exception:
         pass
@@ -47,11 +68,12 @@ app.add_middleware(
 import sys
 sys.path.append(str(Path(__file__).parent))
 
-from api import agents, subagents, websocket, performance, collaboration, agents_config, errors, timeline, chains, agent_config_api, error_analysis, debug_paths, version
+from api import agents, subagents, websocket, performance, collaboration, agents_config, errors, timeline, chains, agent_config_api, error_analysis, debug_paths, version, fortify_routes
 
 # 注册 API 路由
 app.include_router(agents.router, prefix="/api", tags=["agents"])
 app.include_router(errors.router, prefix="/api", tags=["errors"])
+app.include_router(fortify_routes.router, prefix="/api", tags=["fortify"])
 app.include_router(agents_config.router, prefix="/api", tags=["agents-config"])
 app.include_router(subagents.router, prefix="/api", tags=["subagents"])
 app.include_router(websocket.router, tags=["websocket"])

package/dashboard/mechanism_reader.py

@@ -7,7 +7,7 @@ from typing import Dict, Any, List
 
 
 from data.config_reader import get_openclaw_root, normalize_openclaw_agent_id
-from data.session_reader import normalize_sessions_index
+from data.session_reader import normalize_sessions_index, _load_sessions_index_file
 
 
 def get_agent_mechanisms(agent_id: str) -> Dict[str, Any]:
@@ -30,9 +30,8 @@ def get_agent_mechanisms(agent_id: str) -> Dict[str, Any]:
         return result
     
     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:
             return result
         
         # 取最新 session 的机制信息（兼容 sessions.json 顶层或 entries 嵌套）
@@ -129,6 +128,6 @@ def get_agent_mechanisms(agent_id: str) -> Dict[str, Any]:
 
 def get_all_agents_mechanisms() -> List[Dict[str, Any]]:
     """获取所有 Agent 的机制使用情况"""
-    from .config_reader import get_agents_list
+    from data.config_reader import get_agents_list
     agents = get_agents_list()
     return [get_agent_mechanisms(a.get('id', '')) for a in agents if a.get('id')]

package/dashboard/mechanisms.py

@@ -14,14 +14,14 @@ router = APIRouter()
 @router.get("/mechanisms")
 async def get_mechanisms():
     """获取所有 Agent 的机制使用情况"""
-    from data.mechanism_reader import get_all_agents_mechanisms
+    from mechanism_reader import get_all_agents_mechanisms
     return get_all_agents_mechanisms()
 
 
 @router.get("/mechanisms/{agent_id}")
 async def get_agent_mechanisms(agent_id: str):
     """获取单个 Agent 的机制使用情况"""
-    from data.mechanism_reader import get_agent_mechanisms
+    from mechanism_reader import get_agent_mechanisms
     from data.config_reader import get_agent_config
 
     if not get_agent_config(agent_id):

package/dashboard/pytest.ini

@@ -0,0 +1,3 @@
+[pytest]
+markers =
+    benchmark: 可选性能烟测（对负载敏感，默认与主套件一起跑）

package/dashboard/requirements.txt

@@ -4,3 +4,8 @@ pydantic==2.11.7
 python-multipart==0.0.20
 watchdog>=3.0.0
 tzdata
+jsonschema>=4.0.0
+psutil>=5.9.0
+pytest>=7.0.0
+pytest-asyncio>=0.23.0
+httpx>=0.27.0

package/dashboard/status/cache_fp_probe.py

@@ -0,0 +1,40 @@
+"""可选后台线程：周期性对 StatusCache 做 mtime 双验证剔除（RISK-004）。"""
+from __future__ import annotations
+
+import logging
+import threading
+from typing import Optional
+
+_LOG = logging.getLogger("openclaw.fortify.cache_probe")
+
+
+def start_cache_fp_probe_background() -> Optional[threading.Event]:
+    """
+    若 OPENCLAW_CACHE_FP_PROBE_INTERVAL > 0，启动守护线程并返回用于停止的 Event；
+    否则返回 None。
+    """
+    from core.config_fortify import get_fortify_config
+
+    interval = get_fortify_config().cache_fp_probe_interval_sec
+    if interval <= 0:
+        return None
+
+    stop = threading.Event()
+
+    def loop() -> None:
+        from status.status_cache import get_cache
+
+        while not stop.is_set():
+            if stop.wait(timeout=interval):
+                break
+            try:
+                n = get_cache().invalidate_stale_fp_entries()
+                if n:
+                    _LOG.info("cache_fp_probe removed %s stale cache entries", n)
+            except Exception as e:
+                from core.error_handler import record_error
+
+                record_error("unknown", str(e), "cache_fp_probe", exc=e)
+
+    threading.Thread(target=loop, daemon=True, name="openclaw_cache_fp_probe").start()
+    return stop