codebee 0.1.68 → 0.1.69

package/app/core/zentao.py

@@ -1,2465 +1,2465 @@
-# -*- coding: utf-8 -*-
-"""禅道 Bug 自动修复对接：多产品档案 + 排查定责路由 + 转派流转。
-
-数据落盘 <data>/zentao.json（tmp + os.replace 原子写；TUTTI_DATA 环境变量感知）。
-修复走与 /api/tasks 完全相同的链路（store.create_task → store.create_run →
-jobs.enqueue，code 引擎=实现→验证→评审→修复），不自造运行器。调度挂在
-automation._tick（同 publish/auto.fire_due 模式），内部按 interval_minutes 节流。
-
-禅道 REST API v1（开源版 15.x+；请求头 Token: <token>）：
-  POST {base}/api.php/v1/tokens               {account, password} → {token}
-  GET  {base}/api.php/v1/products/{pid}/bugs  分页 {bugs:[...], page, total, limit}
-  GET  {base}/api.php/v1/bugs/{id}            单查（回写前确认状态防谎报）
-  POST {base}/api.php/v1/bugs/{id}/resolve    {resolution, resolvedBuild, comment, assignedTo}
-  PUT  {base}/api.php/v1/bugs/{id}            {assignedTo, comment}（转派/失败说明）
-  GET  {base}/api.php/v1/products|users       产品/账号清单（前端下拉辅助）
-
-地址自适应：根路径 tokens 404 时自动试 {base}/zentao/api.php/v1（官方一键安装包
-默认子路径部署），探测成功缓存在 _APIBASE（key=用户填的原始地址）。
-
-产品档案（product_profiles）：每产品一份 {指派过滤, 严重度, 我方端 our_sides,
-后端/前端仓库(workdir/git_rev/verify_command), repo_hints, 负责人 owners,
-模块路由 module_routes}。老版扁平配置（products+单仓库）load 时自动迁移。
-
-排查（_triage）：模块路由按 bug.module 精确匹配优先 → AI 兜底（triage_ai 开时
-modelhub 单次调用，bookmeta 同款配方）→ unknown。判定 side ∈
-backend | frontend | both | not_ours | unknown。
-
-流转规则（转派目标只认排查结论，与 bug 当前 assignedTo 无关——测试提错人也
-照样改派）：
-  我方端问题    → 建修复任务；成功=合并+resolve(fixed)+报告评论+指回报告人
-  双端/我方一端 → 我方端修完（合并落库）后转派另一端负责人+评论，不 resolve
-  纯对方端问题  → 不建任务，直接转派该端负责人+排查结论评论
-  非我方        → 转派报告人（或 owners.not_ours）+评论；只转派不解决
-  unknown       → 不碰 bug，need_manual + 群通知（下轮扫描 bug 仍激活则重排查）
-  修复任务失败  → 评论尝试记录 + 转派该端负责人（模块路由 account > 端负责人）；
-                  CodeBee 自身崩溃（内部异常）不转派——那是工具故障不是修不动，
-                  只评论+群通知人工；负责人未配置同样只评论。
-                  只评论也绝不盲写 PUT：禅道 PUT /bugs/{id} 缺 assignedTo 会
-                  清空指派人，必须带回当前指派人（#27783 案）。
-
-落库纪律：修复任务落单即带基线（档案配了用档案，没配取工作目录当前 HEAD）
-走任务分支隔离，修完的代码自动提交在任务分支上，对账合并后才算落库；
-resolve 前再验一遍——没基线的任务改动只在工作区，有未提交的已跟踪改动
-不 resolve（转 done_manual 人工收口），auto_merge 关闭同理。
-对账（_reconcile）认任务最新一次运行：人工重试换 run 不影响回写判断。
-
-出网边界（SSRF 防护，_guard_url）：请求目标来自用户自配禅道地址——内网按设计
-放行；强制 http(s)、解析主机并阻断云元数据/链路本地地址、禁跟随重定向。
-"""
-from __future__ import annotations
-
-import base64
-import html as _html
-import hashlib
-import io
-import ipaddress
-import json
-import logging
-import os
-import re
-import socket
-import threading
-import time
-import urllib.error
-import urllib.parse
-import urllib.request
-from datetime import datetime, timedelta
-from pathlib import Path
-
-from . import jobs, operations, paths, settings, store, tlsctx
-
-log = logging.getLogger(__name__)
-
-_LOCK = threading.RLock()
-_SCAN_LOCK = threading.Lock()  # 手动/定时扫描单飞，避免重复拉取和重复认领
-_FILE = paths.DATA_DIR / "zentao.json"
-_STATE = {
-    "config": {},        # 持久配置（_CFG_DEFAULTS）
-    "claims": {},        # str(bug_id) → claim dict（v2：含 triage/tasks）
-    "last_scan": "",
-    "next_scan": "",
-    "last_error": "",
-}
-_LOADED = False
-_BOOT_TIMER = None     # 启动补对账的兜底 Timer（测试里要能拿到并取消）
-
-TOKEN_TTL = 23 * 3600
-HTTP_TIMEOUT = 15
-PAGE_LIMIT = 100
-MAX_BUGS = 500
-RESOLVE_MAX_ATTEMPTS = 3
-RETRY_DELAY_MIN = 30
-INTERVAL_MIN, INTERVAL_MAX = 5, 10080   # 扫描间隔（分钟）：5 分钟 ~ 7 天
-INTERVAL_DEFAULT = 5
-_LEGACY_INTERVAL_HOURS_DEFAULT = 2      # 老配置 interval_hours 缺省值（迁移用）
-
-SIDES = ("backend", "frontend")
-TRIAGE_SIDES = ("backend", "frontend", "both", "not_ours")   # unknown 单列
-SIDE_CN = {"backend": "后端", "frontend": "前端"}
-
-# 图片证据：#27754 案——「功能未对齐，具体如图」被纯文本排查脑补成前端样式问题。
-# 排查/修复前先尽量把图抓到手；抓不到就按守卫留人工，绝不盲判。
-_IMG_REF_RE = re.compile(r"如图|见图|截图|附图|下图|上图|图示|图片|图像|录屏|视频")
-TRIAGE_MAX_IMAGES = 4          # 排查提示词最多随附几张图（请求体与 token 预算）
-IMG_CACHE_TTL = 900            # bug 图片按 bug_id 短缓存（排查+建任务两处共用）
-_IMG_CACHE = {}                # str(bug_id) → (ts, [(mime, b64)])；空结果也缓存防重复下载
-
-_CFG_DEFAULTS = {
-    "base_url": "",
-    "account": "",
-    "password": "",
-    "product_profiles": [],    # 产品档案列表（_norm_profile 形状）
-    "auto_resolve": True,
-    "auto_merge": True,
-    "triage_ai": True,         # 模块路由未命中时用 AI 兜底排查
-    "poll_enabled": False,
-    "interval_minutes": INTERVAL_DEFAULT,
-}
-
-_REPO_DEFAULTS = {"workdir": "", "git_rev": "", "verify_command": ""}
-
-_PROFILE_DEFAULTS = {
-    "product": 0,              # 产品 ID（必填唯一）
-    "assigned_to": "",         # 只认领指派给该账号的 bug；空=不按指派过滤
-    "severity_cap": 0,         # 严重度上限（1 最严重）；0=不限
-    "our_sides": ["backend"],  # 我方端（CodeBee 自动修）；空=纯排查转派不修
-    "repos": {"backend": dict(_REPO_DEFAULTS), "frontend": dict(_REPO_DEFAULTS)},
-    "repo_hints": {"backend": "", "frontend": ""},   # AI 排查时的一句仓库描述
-    "owners": {"backend": "", "frontend": "", "not_ours": ""},
-    "module_routes": [],       # [{module:int, side:backend|frontend|both|not_ours, account:""}]
-}
-
-_UPDATABLE = ("base_url", "account", "password", "product_profiles",
-              "auto_resolve", "auto_merge", "triage_ai",
-              "poll_enabled", "interval_minutes")
-
-
-# ---------------------------------------------------------------- 出网边界（SSRF）
-
-_META_HOSTS = {"metadata.google.internal", "metadata.goog"}
-_NO_REDIRECT = None
-
-
-def _no_redirect_opener():
-    """HTTP 重定向一概不跟：目标必须就是用户配置的那台禅道。"""
-    global _NO_REDIRECT
-    if _NO_REDIRECT is None:
-        class _Stop(urllib.request.HTTPRedirectHandler):
-            def redirect_request(self, req, fp, code, msg, headers, newurl):
-                return None
-        _NO_REDIRECT = urllib.request.build_opener(
-            _Stop(), urllib.request.HTTPSHandler(context=tlsctx.context()))
-    return _NO_REDIRECT
-
-
-def _guard_url(url):
-    """出网前的边界校验：协议白名单 + 主机解析阻断云元数据/链路本地地址。
-
-    私网与环回按设计放行（禅道自建在内网是主场景，配置者即本机用户）。
-    校验失败抛 ZenError。返回原 url。
-    """
-    from urllib.parse import urlparse
-    u = urlparse(url)
-    if u.scheme not in ("http", "https"):
-        raise ZenError("禅道地址只允许 http/https 协议")
-    host = (u.hostname or "").lower()
-    if not host:
-        raise ZenError("禅道地址缺少主机名")
-    if host in _META_HOSTS:
-        raise ZenError("不允许访问云元数据地址")
-    try:
-        infos = socket.getaddrinfo(host, None)
-    except OSError:
-        raise ZenError("禅道主机解析失败：%s" % host)
-    for info in infos:
-        ip = ipaddress.ip_address(info[4][0])
-        if ip.is_link_local or ip.is_reserved and not ip.is_private:
-            raise ZenError("不允许访问链路本地/保留地址：%s" % ip)
-        if ip.is_multicast or ip.is_unspecified:
-            raise ZenError("不允许访问多播/未指定地址：%s" % ip)
-    return url
-
-
-# ---------------------------------------------------------------- 持久化
-
-def _save_locked():
-    _FILE.parent.mkdir(parents=True, exist_ok=True)
-    tmp = _FILE.with_suffix(".tmp")
-    tmp.write_text(json.dumps({"version": 2, **_STATE},
-                              ensure_ascii=False, indent=2), encoding="utf-8")
-    os.replace(str(tmp), str(_FILE))
-
-
-def _normalize_claim(d):
-    """claim 归一：v1 单任务形状（task_id/run_id）→ v2 tasks 数组。"""
-    c = dict(d) if isinstance(d, dict) else {}
-    if isinstance(c.get("tasks"), list) and c["tasks"]:
-        tasks = [dict(t) for t in c["tasks"] if isinstance(t, dict)]
-    else:
-        tasks = [{"side": "backend", "task_id": c.get("task_id") or "",
-                  "run_id": c.get("run_id") or "", "state": "fixing"}] \
-            if (c.get("task_id") or c.get("run_id")) else []
-    c["tasks"] = tasks
-    c.setdefault("triage", {"side": "", "reason": "", "by": "", "account": ""})
-    if not isinstance(c.get("triage"), dict):
-        c["triage"] = {"side": "", "reason": "", "by": "", "account": ""}
-    c.setdefault("note", "")
-    c.setdefault("attempts", 0)
-    return c
-
-
-def _migrate_legacy(cfg):
-    """老扁平配置（products+单仓库）→ product_profiles。就地改写返回。"""
-    if cfg.get("product_profiles"):
-        return cfg
-    products = cfg.get("products") or []
-    if not isinstance(products, list) or not products:
-        return cfg
-    try:
-        pids = [int(p) for p in products if str(p).strip()]
-    except (TypeError, ValueError):
-        return cfg
-    backend = dict(_REPO_DEFAULTS)
-    for k in _REPO_DEFAULTS:
-        backend[k] = str(cfg.get(k) or "")
-    our = ["backend"]
-    if not backend["workdir"]:
-        our = []          # 老配置连工作目录都没配：纯路由
-    for pid in pids:
-        cfg.setdefault("product_profiles", []).append({
-            "product": pid,
-            "assigned_to": str(cfg.get("assigned_to") or ""),
-            "severity_cap": int(cfg.get("severity_cap") or 0),
-            "our_sides": list(our),
-            "repos": {"backend": backend, "frontend": dict(_REPO_DEFAULTS)},
-            "repo_hints": {"backend": "", "frontend": ""},
-            "owners": {"backend": "", "frontend": "", "not_ours": ""},
-            "module_routes": [],
-        })
-    return cfg
-
-
-def load(force=False):
-    """从磁盘加载状态（幂等）。返回 claims 数。测试可重绑 _FILE 后 force=True。"""
-    global _LOADED
-    with _LOCK:
-        if _LOADED and not force:
-            return len(_STATE["claims"])
-        try:
-            data = json.loads(_FILE.read_text(encoding="utf-8"))
-        except Exception:
-            data = {}
-        cfg = data.get("config") if isinstance(data, dict) else None
-        # 老扁平配置先迁移成产品档案（老键 products/workdir/... 不在新 defaults 里，
-        # 必须在按 _CFG_DEFAULTS 过滤之前完成迁移）
-        if isinstance(cfg, dict) and not cfg.get("product_profiles"):
-            cfg = _migrate_legacy(dict(cfg))
-        # 老配置用 interval_hours（小时），新配置改用 interval_minutes（分钟）：
-        # 用户没动过间隔（还是老默认 2 小时）就跟随新默认 5 分钟；真改过的按小时换算保留。
-        if isinstance(cfg, dict) and "interval_minutes" not in cfg \
-                and "interval_hours" in cfg:
-            try:
-                hours = int(cfg.get("interval_hours") or _LEGACY_INTERVAL_HOURS_DEFAULT)
-            except (TypeError, ValueError):
-                hours = _LEGACY_INTERVAL_HOURS_DEFAULT
-            cfg = dict(cfg)
-            mins = INTERVAL_DEFAULT if hours == _LEGACY_INTERVAL_HOURS_DEFAULT else hours * 60
-            cfg["interval_minutes"] = max(INTERVAL_MIN, min(INTERVAL_MAX, mins))
-        merged = dict(_CFG_DEFAULTS)
-        if isinstance(cfg, dict):
-            merged.update({k: v for k, v in cfg.items() if k in _CFG_DEFAULTS})
-        claims = data.get("claims") if isinstance(data, dict) else None
-        _STATE["config"] = merged
-        _STATE["claims"] = {str(k): _normalize_claim(v)
-                            for k, v in (claims or {}).items()} \
-            if isinstance(claims, dict) else {}
-        _STATE["last_scan"] = str(data.get("last_scan") or "") if isinstance(data, dict) else ""
-        _STATE["next_scan"] = str(data.get("next_scan") or "") if isinstance(data, dict) else ""
-        _STATE["last_error"] = str(data.get("last_error") or "") if isinstance(data, dict) else ""
-        _LOADED = True
-        return len(_STATE["claims"])
-
-
-def _ensure_loaded():
-    if not _LOADED:
-        load()
-
-
-def _cfg():
-    cfg = dict(_CFG_DEFAULTS)
-    cfg.update({k: v for k, v in (_STATE.get("config") or {}).items()
-                if k in _CFG_DEFAULTS})
-    return cfg
-
-
-def _profiles(cfg=None):
-    out = []
-    for p in (cfg or _cfg()).get("product_profiles") or []:
-        prof = dict(_PROFILE_DEFAULTS)
-        prof.update({k: v for k, v in (p or {}).items() if k in _PROFILE_DEFAULTS})
-        repos = dict(_PROFILE_DEFAULTS["repos"])
-        for side in SIDES:
-            r = dict(_REPO_DEFAULTS)
-            r.update({k: v for k, v in ((prof.get("repos") or {}).get(side) or {}).items()
-                      if k in _REPO_DEFAULTS})
-            repos[side] = r
-        prof["repos"] = repos
-        hints = dict(_PROFILE_DEFAULTS["repo_hints"])
-        for side in SIDES:
-            hints[side] = str((prof.get("repo_hints") or {}).get(side) or "")
-        prof["repo_hints"] = hints
-        owners = dict(_PROFILE_DEFAULTS["owners"])
-        for k in owners:
-            owners[k] = str((prof.get("owners") or {}).get(k) or "").strip()
-        prof["owners"] = owners
-        prof["our_sides"] = [s for s in SIDES if s in (prof.get("our_sides") or [])]
-        prof["module_routes"] = [r for r in (prof.get("module_routes") or [])
-                                 if isinstance(r, dict)]
-        out.append(prof)
-    return out
-
-
-def _profile_for(cfg, product_id):
-    try:
-        pid = int(product_id or 0)
-    except (TypeError, ValueError):
-        return None
-    for p in _profiles(cfg):
-        if p["product"] == pid:
-            return p
-    return None
-
-
-# ---------------------------------------------------------------- 禅道客户端
-
-class ZenError(Exception):
-    """禅道接口错误（message 人话，可直接展示）。"""
-
-
-_TOKEN = {"v": "", "at": 0.0}
-# 原始地址 → 探测成功的 api 基址（一键安装包常部署在 /zentao 子路径下，自动补探）
-_APIBASE = {}
-# 通道自适应：原始地址 → "rest"（≥15 REST v1）/ "old"（老版 module-method JSON 接口）
-_MODE = {}
-# web 端路由形态：原始地址 → "pathinfo"（伪静态 bug-view-N.html，官方默认）
-# / "get"（index.php?m=bug&f=view&bugID=N）。bug 详情链接用（2026-09-23 真机实证
-# 伪静态部署对 GET 式链接不路由）。
-_WEBSTYLE = {}
-_OLD = {"api": "", "sid": "", "at": 0.0}
-_OLD_FORM = {"form": ""}      # 命中的密码形态 "plain" | "md5chain"，缓存避免重复试
-_REST_LAST_ERR = {"msg": ""}
-_OLD_TTL = 20 * 60
-
-
-def _reset_token():
-    _TOKEN["v"] = ""
-    _TOKEN["at"] = 0.0
-
-
-def resolved_base_url(base_url):
-    """探测成功后的有效地基（含自动补出的子路径）；没探测过返回空串。"""
-    b = str(base_url or "").strip().rstrip("/")
-    hit = _APIBASE.get(b) or (_OLD.get("api") if _MODE.get(b) == "old" else "")
-    if not hit:
-        return ""
-    return re.sub(r"/api\.php/v1$", "", hit)
-
-
-def probe_web_style(base_url):
-    """探测禅道 web 端路由形态（bug 详情链接用）：GET web 根（裸根 + 常见 /zentao
-    子路径都试），看登录跳转里是伪静态路由（user-login-xxx.html）还是 GET 式
-    （index.php?m=user）。尽力而为：任何网络/守卫异常回空串，绝不影响登录与
-    扫描主流程；探出即缓存。"""
-    try:
-        b = _raw_base(base_url)
-    except ZenError:
-        return ""
-    if _WEBSTYLE.get(b):
-        return _WEBSTYLE[b]
-    style = ""
-    for root in (b + "/", b + "/zentao/"):
-        body = ""
-        try:
-            req = urllib.request.Request(_guard_url(root), headers={"Accept": "text/html"})
-            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
-                body = (r.read(65536) or b"").decode("utf-8", "replace")
-        except Exception:
-            body = ""
-        if "user-login-" in body or re.search(r"[a-z]+-view-\d+[^\"]*\.html", body):
-            style = "pathinfo"
-        elif "index.php?m=" in body:
-            style = "get"
-        if style:
-            break
-    if style:
-        _WEBSTYLE[b] = style
-    return style
-
-
-def _raw_base(base_url):
-    b = str(base_url or "").strip().rstrip("/")
-    if not b:
-        raise ZenError("禅道地址未配置")
-    if not b.startswith(("http://", "https://")):
-        raise ZenError("禅道地址必须以 http:// 或 https:// 开头")
-    return b
-
-
-def _api_base(base_url):
-    b = _raw_base(base_url)
-    hit = _APIBASE.get(b)
-    if hit:
-        return hit
-    return b + "/api.php/v1"
-
-
-def _base_candidates(b):
-    """api 基址候选：用户填的根路径 → 官方一键安装包常见的 /zentao 子路径。"""
-    out, seen = [], set()
-    for c in (b + "/api.php/v1", b + "/zentao/api.php/v1"):
-        if c not in seen:
-            seen.add(c)
-            out.append(c)
-    return out
-
-
-def _rest_login(base_url, account, password):
-    """REST v1 通道取 token。成功返回 token（缓存基址）；
-    通道不可用（404 / 只认应用 code 的 200 信封）返回 None；
-    账密被 REST 明确拒绝（401/403）或网络/守卫失败抛 ZenError。"""
-    b = _raw_base(base_url)
-    for api in _base_candidates(b):
-        url = _guard_url(api + "/tokens")
-        body = json.dumps({"account": str(account or ""), "password": str(password or "")},
-                          ensure_ascii=False).encode("utf-8")
-        req = urllib.request.Request(url, data=body, method="POST", headers={
-            "Content-Type": "application/json", "Accept": "application/json"})
-        try:
-            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
-                data = json.loads((r.read() or b"{}").decode("utf-8", "replace"))
-        except urllib.error.HTTPError as e:
-            try:
-                detail = json.loads((e.read() or b"").decode("utf-8", "replace"))
-                msg = (detail.get("error") or "") if isinstance(detail, dict) else ""
-            except Exception:
-                msg = ""
-            if e.code in (401, 403):
-                raise ZenError("禅道账号或密码不对（%s）%s" % (e.code, msg))
-            if e.code == 404:
-                _REST_LAST_ERR["msg"] = "REST 接口不存在（404）"
-                continue        # 换下一个候选基址
-            raise ZenError("禅道返回 %s：%s" % (e.code, msg or "获取令牌失败"))
-        except ZenError:
-            raise
-        except Exception as e:
-            raise ZenError("连不上禅道（%s）——请检查地址与网络" % e)
-        tok = ""
-        if isinstance(data, dict):
-            tok = str(data.get("token") or "")
-            if not tok and isinstance(data.get("data"), dict):
-                tok = str(data["data"].get("token") or "")
-        if tok:
-            _APIBASE[b] = api
-            _TOKEN["v"] = tok
-            _TOKEN["at"] = time.time()
-            probe_web_style(b)      # bug 链接形态探测，失败不影响登录
-            return tok
-        # HTTP 200 但没有 token：REST 在但不认账密（如只认应用 code 的部署）
-        _REST_LAST_ERR["msg"] = "REST 接口不认账密（%s）" % (
-            str(data.get("errmsg") or data.get("error") or "响应里没有 token"))
-    return None
-
-
-def _token(cfg, force=False):
-    if force or not _TOKEN["v"] or time.time() - _TOKEN["at"] > TOKEN_TTL:
-        tok = _rest_login(cfg.get("base_url"), cfg.get("account"), cfg.get("password"))
-        if not tok:
-            raise ZenError(_REST_LAST_ERR.get("msg") or "REST 通道不可用")
-        return tok
-    return _TOKEN["v"]
-
-
-def _api(method, path, cfg=None, body=None):
-    """调禅道 API。返回 dict（非 dict 响应返回 {}）。401/403 自动重取 token 重试一次。"""
-    c = cfg or _cfg()
-    for attempt in (1, 2):
-        headers = {"Accept": "application/json", "Token": _token(c, force=(attempt == 2))}
-        url = _guard_url(_api_base(c.get("base_url")) + path)
-        data = None
-        if body is not None:
-            data = json.dumps(body, ensure_ascii=False).encode("utf-8")
-            headers["Content-Type"] = "application/json"
-        req = urllib.request.Request(url, data=data, method=method, headers=headers)
-        try:
-            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
-                out = json.loads((r.read() or b"{}").decode("utf-8", "replace"))
-            return out if isinstance(out, dict) else {"_list": out}
-        except urllib.error.HTTPError as e:
-            raw = ""
-            try:
-                raw = (e.read() or b"").decode("utf-8", "replace")
-            except Exception:
-                pass
-            detail = ""
-            try:
-                d = json.loads(raw)
-                if isinstance(d, dict):
-                    detail = str(d.get("error") or d.get("message") or "")
-            except Exception:
-                pass
-            if e.code in (401, 403) and attempt == 1:
-                _reset_token()
-                continue
-            raise ZenError("禅道接口 %s（%s）%s" % (path, e.code, detail or "调用失败"))
-        except ZenError:
-            raise
-        except Exception as e:
-            raise ZenError("连不上禅道（%s）" % e)
-    raise ZenError("禅道认证失败（token 两次获取后仍被拒绝，检查账号权限）")
-
-
-# ---------------------------------------------------------------- 老版 JSON 接口适配
-# 禅道 15 以前的经典入口：/zentao/api-getsessionid.json 取会话 → user-login.json
-# 账密登录（Cookie zentaosid）→ {module}-{method}-{参数}.json 调业务。响应是双层
-# 信封 {"status","data":"<json 字符串>"}。有的部署 REST 只认应用 code，账密只能
-# 走这条通道，故在 _call 里自动探测分路。REST 路径在这里翻译成老接口形态。
-
-def _old_parse(raw):
-    """老信封解析。非 JSON（多半是登录重定向 HTML）返回 None = 会话死/不可用。"""
-    try:
-        outer = json.loads(raw.decode("utf-8", "replace"))
-    except Exception:
-        return None
-    if isinstance(outer, dict) and isinstance(outer.get("data"), str):
-        try:
-            outer["data"] = json.loads(outer["data"])
-        except Exception:
-            pass
-    return outer if isinstance(outer, dict) else None
-
-
-def _old_fail(outer):
-    """status=failed / result=fail → 人话错误文本；成功返回空串。"""
-    if not isinstance(outer, dict):
-        return ""
-    if outer.get("status") == "failed" or outer.get("result") == "fail":
-        return str(outer.get("reason") or outer.get("message")
-                   or outer.get("error") or "调用失败")
-    return ""
-
-
-def _old_get(api, path):
-    url = _guard_url(api + path)
-    req = urllib.request.Request(url, headers={"Accept": "application/json"})
-    with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
-        return r.read() or b""
-
-
-def _old_post(api, path, form):
-    url = _guard_url(api + path)
-    data = urllib.parse.urlencode(form or {}).encode("utf-8")
-    req = urllib.request.Request(url, data=data, method="POST", headers={
-        "Content-Type": "application/x-www-form-urlencoded",
-        "Accept": "application/json"})
-    with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
-        return r.read() or b""
-
-
-def _old_session(api):
-    """取新会话。返回 (sid, rand)；通道不可用返回 None。"""
-    try:
-        raw = _old_get(api, "/api-getsessionid.json")
-    except urllib.error.HTTPError:
-        return None
-    except ZenError:
-        raise
-    except Exception as e:
-        raise ZenError("连不上禅道（%s）——请检查地址与网络" % e)
-    outer = _old_parse(raw)
-    if not outer or not isinstance(outer.get("data"), dict):
-        return None
-    d = outer["data"]
-    return str(d.get("sessionID") or ""), str(d.get("rand") or "")
-
-
-def _old_probe_authed(api, sid):
-    """登录成功的兜底判据：带会话调需登录端点拿得到 JSON 信封（未登录会被
-    重定向到登录页 HTML，解析出来是 None）。"""
-    try:
-        raw = _old_get(api, "/my-index.json?zentaosid=" + sid)
-    except Exception:
-        return False
-    return _old_parse(raw) is not None
-
-
-def _old_identify(api, sid, rand, account, password, form):
-    """老接口账密登录。返回 (ok, 人话错误)。form=plain|md5chain。"""
-    pw = str(password or "")
-    if form == "md5chain":
-        pw = hashlib.md5((hashlib.md5(pw.encode("utf-8")).hexdigest()
-                          + str(rand)).encode("utf-8")).hexdigest()
-    try:
-        raw = _old_post(api, "/user-login.json?zentaosid=" + sid,
-                        {"account": str(account or ""), "password": pw})
-    except Exception as e:
-        return False, "连不上禅道（%s）" % e
-    outer = _old_parse(raw)
-    if outer is None:
-        return False, "登录响应不可识别"
-    err = _old_fail(outer)
-    if err:
-        return False, err
-    # 成功响应形状各版本不一：user 可能在 data 里（老）或顶层（实测某老版部署）
-    d = outer.get("data")
-    if (isinstance(d, dict) and isinstance(d.get("user"), dict)) \
-            or isinstance(outer.get("user"), dict) or _old_probe_authed(api, sid):
-        return True, ""
-    return False, "登录未被接受（账号或密码不对，或账号被锁）"
-
-
-def _old_login(cfg, rest_err=""):
-    """老接口通道登录（带 /zentao 子路径候选）。成功置 _OLD 缓存与密码形态。"""
-    b = _raw_base(cfg.get("base_url"))
-    account, password = cfg.get("account"), cfg.get("password")
-    cands, seen = [], set()
-    for c in (b + "/zentao", b):
-        if c not in seen:
-            seen.add(c)
-            cands.append(c)
-    last = ""
-    for api in cands:
-        sess = _old_session(api)
-        if not sess:
-            continue
-        sid, rand = sess
-        forms = [_OLD_FORM["form"] or "plain", "md5chain", "plain"]
-        tried = set()
-        for form in [f for f in forms if not (f in tried or tried.add(f))]:
-            ok, err = _old_identify(api, sid, rand, account, password, form)
-            if ok:
-                _OLD.update(api=api, sid=sid, at=time.time())
-                _OLD_FORM["form"] = form
-                _MODE[b] = "old"
-                probe_web_style(b)  # bug 链接形态探测，失败不影响登录
-                return
-            last = err
-        break       # 同一台服务，账密结果与子路径无关，别再烧尝试次数
-    if not last:
-        raise ZenError("禅道老版接口不可用（会话接口无响应——地址若是子目录部署要带上子目录）"
-                       + ("；%s" % rest_err if rest_err else ""))
-    if "账号" in last or "密码" in last or "锁定" in last:
-        raise ZenError(last)
-    raise ZenError("禅道老接口登录失败：%s" % last)
-
-
-def _old_call(method, path, cfg, body=None):
-    """老接口执行：翻译 REST 路径 → module-method.json，返回 REST 同形状的 dict。"""
-    b = _raw_base(cfg.get("base_url"))
-    if _MODE.get(b) != "old" or not _OLD["sid"] or time.time() - _OLD["at"] > _OLD_TTL:
-        _old_login(cfg)
-    parts = urllib.parse.urlsplit(path)
-    q = dict(urllib.parse.parse_qsl(parts.query))
-    p = parts.path
-    for attempt in (1, 2):
-        try:
-            return _old_route(_OLD["api"], method, p, q, body)
-        except _OldSessionDead:
-            if attempt == 2:
-                break
-            _OLD["sid"] = ""
-            _old_login(cfg)
-        except urllib.error.HTTPError as e:
-            raise ZenError("禅道老接口 %s（%s）调用失败" % (p, e.code))
-    raise ZenError("禅道会话两次登录后仍失效（检查账号权限）")
-
-
-class _OldSessionDead(Exception):
-    pass
-
-
-def _pairs_to_list(pairs, id_key, name_key):
-    """{id/name: 值} 形态的 pairs → [{id,name}] 列表（值可能为串或对象）。"""
-    out = []
-    for k, v in (pairs or {}).items():
-        if isinstance(v, dict):
-            item = dict(v)
-            item.setdefault(id_key, k)
-        else:
-            item = {id_key: k, name_key: v}
-        out.append(item)
-    return out
-
-
-def _module_items(d):
-    """老版模块树形状兼容（2026-09-21 真机反馈「响应形状不认识」）：
-    sons/modules 数组、{id: {name,...}} 字典、children/sons 嵌套树，统一
-    递归展开成 [{id,name}] 平铺清单（id=0 是树根容器，跳过）；全数字键的
-    {id: 名字} 扁平映射也认（部分版本的模块摘要形状）。"""
-    out = []
-
-    def walk(node):
-        if isinstance(node, list):
-            for x in node:
-                walk(x)
-        elif isinstance(node, dict):
-            if node.get("id") not in (None, "", 0):
-                out.append(node)
-                for k in ("children", "sons"):
-                    if isinstance(node.get(k), (list, dict)):
-                        walk(node[k])
-            else:
-                if node and all(str(k).isdigit() for k in node.keys()):
-                    flat = [{"id": k, "name": str(v)}
-                            for k, v in node.items() if str(v or "").strip()]
-                    if flat:
-                        out.extend(flat)
-                        return
-                for v in node.values():
-                    walk(v)
-
-    if isinstance(d, dict):
-        # tree 优先：真机 tree=带 children 的完整嵌套树，sons 只有顶层扁平表
-        # （走 sons 会漏子模块）
-        walk(d.get("tree") or d.get("sons") or d.get("modules") or d)
-    elif isinstance(d, list):
-        walk(d)
-    return out
-
-
-def _old_route(api, method, p, q, body):
-    """路径翻译 + 请求 + 归一。会话死抛 _OldSessionDead。"""
-
-    def go(raw, translator=None):
-        outer = _old_parse(raw)
-        if outer is None:
-            txt = raw.decode("utf-8", "replace")
-            if "user-login" in txt:
-                raise _OldSessionDead()     # 登录重定向 = 会话死
-            # 其余 HTML 是 js::locate/alert 回包（老禅道 POST 动作成功就回这种）
-            return {"_js": True}
-        err = _old_fail(outer)
-        if err:
-            raise ZenError("禅道老接口%s：%s" % (translator or "", err))
-        return outer.get("data")
-
-    def get(path):
-        try:
-            raw = _old_get(api, path + ("&" if "?" in path else "?")
-                           + "zentaosid=" + _OLD["sid"])
-        except urllib.error.HTTPError as e:
-            raw = e.read() or b""
-            if _old_parse(raw) is None:
-                raise _OldSessionDead()     # 非 JSON：多半被重定向到登录页
-        return raw
-
-    def post(path, form):
-        try:
-            raw = _old_post(api, path + ("&" if "?" in path else "?")
-                            + "zentaosid=" + _OLD["sid"], form)
-        except urllib.error.HTTPError as e:
-            raw = e.read() or b""
-            if e.code in (401, 403) or _old_parse(raw) is None:
-                raise _OldSessionDead()
-        return raw
-
-    m = re.match(r"^/products/(\d+)/bugs$", p)
-    if m and method == "GET":
-        pid = m.group(1)
-        page = max(1, int(q.get("page") or 1))
-        out, total, seen_ids = [], 0, set()
-        # 翻页用路径参数形态（实测部分老版不认 query 翻页参数）：带 branch 段；
-        # 若 recPerPage 没生效（老版本无 branch 段会错位解析）回落 query 形态。
-        branch_form = True
-        for pg in range(page, page + 6):     # 去重兜底：翻页参数不被认时不会死循环
-            if branch_form:
-                pathq = "/bug-browse-%s-0-unclosed-0-id_desc-0-%d-%d.json" % (pid, PAGE_LIMIT, pg)
-            else:
-                pathq = "/bug-browse-%s.json?browseType=unclosed&orderBy=id_desc" \
-                        "&recPerPage=%d&pagerID=%d" % (pid, PAGE_LIMIT, pg)
-            d = go(get(pathq), "（拉 bug 列表）")
-            if branch_form and isinstance(d, dict):
-                try:
-                    got = int((d.get("pager") or {}).get("recPerPage") or 0)
-                except (TypeError, ValueError):
-                    got = 0
-                if got != PAGE_LIMIT:
-                    branch_form = False
-                    continue                 # 形态没对上：换 query 形态重拉本页
-            if not isinstance(d, dict):
-                break
-            bugs = d.get("bugs") if isinstance(d.get("bugs"), list) else []
-            fresh = [x for x in bugs if isinstance(x, dict)
-                     and str(x.get("id")) not in seen_ids]
-            for x in fresh:
-                seen_ids.add(str(x.get("id")))
-            out.extend(fresh)
-            try:
-                total = int((d.get("pager") or {}).get("recTotal") or 0)
-            except (TypeError, ValueError, AttributeError):
-                total = 0
-            if not fresh or (total and len(out) >= total) or len(out) >= MAX_BUGS:
-                break
-        return {"bugs": out[:MAX_BUGS], "total": total or len(out)}
-
-    m = re.match(r"^/products/(\d+)/modules$", p)
-    if m and method == "GET":
-        # 2026-09-22 真机实探：该版模块行 type=story，viewType=module 查出空树
-        # （status 仍 success）——「响应形状不认识」的真根因。改按 bug 视图拉
-        # （bug 表单模块下拉用的就是这棵树），story/case 兜底。
-        pid = m.group(1)
-        last, saw_js = None, False
-        for vt in ("bug", "story", "case"):
-            d = go(get("/tree-browse-%s-%s.json" % (pid, vt)), "（拉模块清单）")
-            last = d
-            if isinstance(d, dict) and d.get("_js"):
-                saw_js = True            # 接口回了页面脚本：换视图再试
-                continue
-            items = [x for x in _module_items(d) if isinstance(x, dict) and x.get("id")]
-            if items:
-                return {"_list": items}
-        if saw_js:
-            raise ZenError("禅道老接口（拉模块清单）：回了页面脚本而不是数据"
-                           "（模块树接口不在或被重定向）——请在禅道产品视图 URL 里查模块 ID 手工填写")
-        keys = ",".join(sorted(str(k) for k in last.keys())) \
-            if isinstance(last, dict) else type(last).__name__
-        raise ZenError("该产品模块树为空（bug/story/case 三个视图都没拉到模块，"
-                       "响应顶层键：%s）——请到禅道确认产品下建过模块，或手工填模块 ID" % keys)
-
-    if p == "/products" and method == "GET":
-        d = go(get("/api-getmodel-product-getpairs.json"), "（拉产品清单）")
-        prods = d.get("products") if isinstance(d, dict) else None
-        if isinstance(prods, list):     # 部分老版回数组形状 [{id,name,...}]
-            return {"_list": [x for x in prods if isinstance(x, dict) and x.get("id")]}
-        pairs = prods if isinstance(prods, dict) else d
-        return {"_list": _pairs_to_list(pairs, "id", "name")}
-
-    if p == "/users" and method == "GET":
-        d = go(get("/api-getmodel-user-getpairs.json"), "（拉账号清单）")
-        return {"_list": _pairs_to_list(d, "account", "realname")}
-
-    m = re.match(r"^/bugs/(\d+)$", p)
-    if m and method == "GET":
-        d = go(get("/bug-view-%s.json" % m.group(1)), "（查 bug）")
-        bug = d.get("bug") if isinstance(d, dict) and isinstance(d.get("bug"), dict) else d
-        # 老接口的 files 附件在 data 顶层，不在 bug 里：合并进 bug dict（REST 形状兼容）
-        if isinstance(bug, dict) and isinstance(d, dict) \
-                and isinstance(d.get("files"), (dict, list)):
-            bug = dict(bug)
-            bug["files"] = d["files"]
-        return bug
-
-    m = re.match(r"^/bugs/(\d+)/resolve$", p)
-    if m and method == "POST":
-        form = {k: str(v or "") for k, v in (body or {}).items()}
-        go(post("/bug-resolve-%s.json" % m.group(1), form), "（resolve）")
-        return {"ok": True}
-
-    m = re.match(r"^/bugs/(\d+)$", p)
-    if m and method == "PUT":
-        form = {k: str(v or "") for k, v in (body or {}).items() if k in ("assignedTo", "comment")}
-        go(post("/bug-assignTo-%s.json" % m.group(1), form), "（转派）")
-        return {"ok": True}
-
-    raise ZenError("禅道老接口不认识该调用：%s %s（请升级禅道到 ≥15 用 REST 接口）"
-                   % (method, p))
-
-
-def _call(method, path, cfg=None, body=None):
-    """统一入口：REST v1 优先；通道探明后按模式分发（账密不被 REST 接受时自动
-    切老版 JSON 接口）。所有业务调用都走这里，拿到的都是 REST 形状。"""
-    c = cfg or _cfg()
-    b = _raw_base(c.get("base_url"))
-    mode = _MODE.get(b)
-    if not mode:
-        tok = _rest_login(b, c.get("account"), c.get("password"))
-        if tok:
-            _MODE[b] = mode = "rest"
-        else:
-            _old_login(c, _REST_LAST_ERR.get("msg", ""))
-            mode = "old"
-    if mode == "rest":
-        return _api(method, path, cfg=c, body=body)
-    return _old_call(method, path, c, body=body)
-
-
-def _acct(v):
-    """assignedTo/openedBy 兼容：新版是 {account,...} 用户对象，老版可能是
-    [账号, 姓名] 数组或账号串。"""
-    if isinstance(v, dict):
-        return str(v.get("account") or "").strip()
-    if isinstance(v, (list, tuple)) and v:
-        return str(v[0] or "").strip()
-    return str(v or "").strip()
-
-
-def _strip_html(raw):
-    """steps 字段是富文本 HTML：剥标签转纯文本。"""
-    txt = str(raw or "")
-    txt = re.sub(r"(?is)<(script|style).*?>.*?</\1>", " ", txt)
-    txt = re.sub(r"(?i)<br\s*/?>|</p>|</li>|</div>|</tr>", "\n", txt)
-    txt = re.sub(r"<[^>]+>", " ", txt)
-    txt = _html.unescape(txt)
-    txt = re.sub(r"[ \t\r]+", " ", txt)
-    txt = re.sub(r"\n\s*\n+", "\n", txt)
-    return txt.strip()
-
-
-def _severity(bug):
-    try:
-        return int(bug.get("severity") or 0)
-    except (TypeError, ValueError):
-        return 0
-
-
-# ---------------------------------------------------------------- bug 图片证据
-
-_IMG_MAGIC = ((b"\x89PNG", "image/png"), (b"\xff\xd8\xff", "image/jpeg"),
-              (b"GIF8", "image/gif"))
-
-
-def _img_mime(raw):
-    """按魔数判图片类型；非图返回空串（下载到错误页 HTML 时靠它拦住）。"""
-    raw = bytes(raw or b"")
-    for pre, mime in _IMG_MAGIC:
-        if raw.startswith(pre):
-            return mime
-    if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP":
-        return "image/webp"
-    return ""
-
-
-def _img_ext_mime(ext):
-    return {"png": "image/png", "jpg": "image/jpeg", "jpeg": "image/jpeg",
-            "gif": "image/gif", "webp": "image/webp", "bmp": "image/bmp"}.get(
-        str(ext or "").strip().lstrip(".").lower(), "")
-
-
-def _img_srcs(steps_html):
-    """steps 富文本里的 <img src> 候选（含 data: URI）。"""
-    return re.findall(r'(?is)<img\b[^>]*?\bsrc\s*=\s*["\']([^"\']+)["\']',
-                      str(steps_html or ""))
-
-
-def _bug_mentions_images(bug):
-    """描述是否指向图片证据：steps 带 <img>、文字提「如图/截图…」、或附件含图片。"""
-    html = str(bug.get("steps") or "")
-    if "<img" in html.lower():
-        return True
-    text = str(bug.get("title") or "") + "\n" + _strip_html(html)
-    if _IMG_REF_RE.search(text):
-        return True
-    files = bug.get("files")
-    vals = files.values() if isinstance(files, dict) else (files if isinstance(files, list) else [])
-    for e in vals:
-        if isinstance(e, dict) and _img_ext_mime(
-                e.get("extension") or str(e.get("title") or "").rsplit(".", 1)[-1]):
-            return True
-    return False
-
-
-def _decode_data_uri(src):
-    m = re.match(r"(?is)^data:(image/[\w.+-]+);base64,(.+)$", src.strip())
-    if not m:
-        return "", b""
-    try:
-        raw = base64.b64decode(m.group(2))
-    except Exception:
-        return "", b""
-    mime = _img_mime(raw) or m.group(1).lower()
-    return mime, raw
-
-
-def _download_same_origin(src, cfg):
-    """下载禅道站内图片字节：仅同源（host:port 与配置的禅道一致），带会话多形态
-    尝试（zentaosid 查询参——新老两通道都认；再裸 GET 兜底），魔数校验。
-    非 200 / 非图 / 网络失败返回 b""。"""
-    try:
-        base = _raw_base(cfg.get("base_url"))
-    except ZenError:
-        return b""
-    url = src if "://" in str(src) else base + "/" + str(src).lstrip("/")
-    try:
-        u, bu = urllib.parse.urlsplit(url), urllib.parse.urlsplit(base)
-    except ValueError:
-        return b""
-    if (u.scheme or bu.scheme, u.netloc) != (bu.scheme, bu.netloc):
-        return b""          # 第三方图床一律不取（SSRF 边界外只放行同源）
-    url = _guard_url("%s://%s%s" % (bu.scheme, bu.netloc, u.path or "/")
-                     + (("?" + u.query) if u.query else ""))
-    tries = []
-    for sid in (str(_OLD.get("sid") or ""), str(_TOKEN.get("v") or "")):
-        if sid:
-            tries.append(url + ("&" if "?" in url else "?") + "zentaosid=" + sid)
-    tries.append(url)
-    for t in tries:
-        try:
-            req = urllib.request.Request(t, headers={"Accept": "*/*"})
-            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
-                if r.status != 200:
-                    continue
-                raw = r.read(20 * 1024 * 1024)
-        except Exception:
-            continue
-        if _img_mime(raw):
-            return raw
-    return b""
-
-
-def _file_entry_urls(fid, entry):
-    """详情 files 条目 → 下载候选 URL（entry 自带地址优先，再猜经典路由）。"""
-    out = []
-    if isinstance(entry, dict):
-        for k in ("download", "url", "href"):
-            v = str(entry.get(k) or "").strip()
-            if v.startswith(("http://", "https://", "/")):
-                out.append(v)
-        ext = _img_ext_mime(entry.get("extension")
-                            or str(entry.get("title") or "").rsplit(".", 1)[-1])
-    else:
-        ext = ""
-    if ext:
-        ext = ext.split("/", 1)[1]
-        for route in ("file-read-%s.%s", "file-download-%s.%s"):
-            out.append("/" + route % (fid, ext))
-    return out
-
-
-def _prep_image_pairs(pairs):
-    """[(mime, 原始字节)] → [(mime, b64)]。缩放规则与内置智能体同款（长边 1568、
-    超大转 JPEG q85、单张 5MB 剔除）；就地实现不跨层 import（架构 L2→L1 反向禁）。"""
-    try:
-        from PIL import Image
-    except Exception:
-        Image = None
-    out = []
-    for mime, raw in pairs or []:
-        data = raw
-        if Image is not None:
-            try:
-                im = Image.open(io.BytesIO(raw))
-                im.load()
-                w, h = im.size
-                edge = max(w, h, 1)
-                if edge > 1568:
-                    r = 1568 / float(edge)
-                    im = im.resize((max(1, round(w * r)), max(1, round(h * r))),
-                                   Image.LANCZOS)
-                fmt = {"image/png": "PNG", "image/jpeg": "JPEG",
-                       "image/gif": "GIF", "image/webp": "WEBP"}.get(mime, "PNG")
-                buf = io.BytesIO()
-                im.save(buf, format=fmt)
-                data = buf.getvalue()
-            except Exception:
-                data = raw                # 解码失败退回原始字节
-            if mime != "image/jpeg" and len(data) > 3500 * 1024:
-                try:                      # 仍过大：转 JPEG q85（RGB 拍平 alpha）
-                    im = Image.open(io.BytesIO(data)).convert("RGB")
-                    buf = io.BytesIO()
-                    im.save(buf, format="JPEG", quality=85)
-                    data, mime = buf.getvalue(), "image/jpeg"
-                except Exception:
-                    pass
-        if len(data) > 5 * 1024 * 1024:
-            continue
-        out.append((mime, base64.b64encode(data).decode("ascii")))
-    return out
-
-
-def _bug_images(bug_id, steps_html, cfg=None):
-    """bug 关联图片 → [(mime, b64)]：steps 内嵌 <img>（含 data: URI）优先，
-    读不到再看详情 files 附件。下载带禅道会话、魔数校验、缩放管线同内置智能体。
-    拿不到返回 []（也缓存，15 分钟内不重复下载）。"""
-    key = str(bug_id or "")
-    now = time.time()
-    hit = _IMG_CACHE.get(key)
-    if hit and now - hit[0] < IMG_CACHE_TTL:
-        return hit[1]
-    cfg = cfg or _cfg()
-    pairs = []
-    for src in _img_srcs(steps_html):
-        s = src.strip()
-        if s.lower().startswith("data:"):
-            mime, raw = _decode_data_uri(s)
-            if raw:
-                pairs.append((mime or "image/png", raw))
-        elif not s.lower().startswith("javascript:"):
-            raw = _download_same_origin(s, cfg)
-            if raw:
-                pairs.append((_img_mime(raw), raw))
-    if not pairs and key:
-        try:
-            detail = _call("GET", "/bugs/%s" % key, cfg=cfg)
-            files = detail.get("files") if isinstance(detail, dict) else None
-            entries = (list(files.items()) if isinstance(files, dict)
-                       else [(str(i), e) for i, e in enumerate(files or [])])
-            for fid, entry in entries:
-                mime = _img_ext_mime(
-                    (entry.get("extension") if isinstance(entry, dict) else "")
-                    or str(entry.get("title") if isinstance(entry, dict) else "").rsplit(".", 1)[-1])
-                if not mime:
-                    continue          # 日志/压缩包等非图附件不取
-                for u in _file_entry_urls(fid, entry):
-                    raw = _download_same_origin(u, cfg)
-                    if raw:
-                        pairs.append((_img_mime(raw), raw))
-                        break
-        except Exception:
-            log.debug("zentao: bug 附件图片获取失败", exc_info=True)
-    imgs = _prep_image_pairs(pairs[:TRIAGE_MAX_IMAGES])
-    if len(_IMG_CACHE) > 128:
-        for k in sorted(_IMG_CACHE, key=lambda k: _IMG_CACHE[k][0])[:64]:
-            _IMG_CACHE.pop(k, None)
-    _IMG_CACHE[key] = (now, imgs)
-    return imgs
-
-
-def _save_bug_attachments(workdir, bug_id, imgs):
-    """bug 截图落工作目录 _attachments/（create_task 的 dict 清单形态直用）。
-    返回 [{name,size,mime,path}]；失败返回 []。"""
-    if not imgs:
-        return []
-    try:
-        adir = Path(workdir) / "_attachments"
-        adir.mkdir(parents=True, exist_ok=True)
-        ext = {"image/png": ".png", "image/jpeg": ".jpg", "image/gif": ".gif",
-               "image/webp": ".webp"}
-        items = []
-        for i, (mime, b64) in enumerate(imgs, 1):
-            name = "zenbug%s-%d%s" % (bug_id, i, ext.get(mime, ".png"))
-            data = base64.b64decode(b64)
-            (adir / name).write_bytes(data)
-            items.append({"name": name, "size": len(data), "mime": mime,
-                          "path": "_attachments/" + name})
-        return items
-    except Exception:
-        log.debug("zentao: bug 截图落附件失败", exc_info=True)
-        return []
-
-
-def _claimable(bug, profile):
-    """认领过滤：active + 指派 + 严重度（产品由档案本身界定）。"""
-    if str(bug.get("status") or "") != "active":
-        return False
-    assigned = str(profile.get("assigned_to") or "").strip()
-    cap = int(profile.get("severity_cap") or 0)
-    if assigned and _acct(bug.get("assignedTo")) != assigned:
-        return False
-    if cap and _severity(bug) > cap:
-        return False
-    return True
-
-
-def list_bugs(cfg, product_id, profile=None):
-    """拉一个产品下的激活 bug（分页，总量封顶 MAX_BUGS）。
-
-    REST v1 支持时把状态/指派人过滤下推给禅道，减少无关 bug 的分页传输；
-    本地的 _claimable 仍保留，兼容老接口或禅道忽略未知查询参数的情况。
-    """
-    out = []
-    for page in range(1, 6):
-        query = {"page": page, "limit": PAGE_LIMIT, "status": "active"}
-        assigned = str((profile or {}).get("assigned_to") or "").strip()
-        if assigned:
-            query["assignedTo"] = assigned
-        d = _call("GET", "/products/%s/bugs?%s" % (
-            product_id, urllib.parse.urlencode(query)), cfg=cfg)
-        bugs = d.get("bugs") if isinstance(d, dict) else None
-        if not isinstance(bugs, list):
-            raise ZenError("禅道 bug 列表响应形状不对（预期 bugs 数组）")
-        out.extend(b for b in bugs if isinstance(b, dict))
-        total = 0
-        try:
-            total = int(d.get("total") or 0)
-        except (TypeError, ValueError):
-            pass
-        if len(out) >= MAX_BUGS or (total and len(out) >= total) or len(bugs) < PAGE_LIMIT:
-            break
-    return out[:MAX_BUGS]
-
-
-def weekly_brief(limit=12):
-    """本周工作素材（工作汇报起草注入用，2026-09-23 Weekly Report Generator 借鉴
-    ——「从工作系统取数入素材」）：分配给我的活跃任务 + 指派给我的活跃 Bug 摘要。
-    尽力而为：未配置/不可达/响应形状不对一律返回空串，绝不阻塞汇报起草。
-    多产品档案逐个聚合（同 _scan 口径），条目总量封顶 limit*2。"""
-    lines = []
-    try:
-        cfg = _cfg()
-        for profile in _profiles(cfg):
-            assigned = str(profile.get("assigned_to") or "").strip()
-            if not assigned:
-                continue
-            try:
-                d = _call("GET", "/tasks?%s" % urllib.parse.urlencode(
-                    {"assignedTo": assigned, "limit": limit}), cfg=cfg)
-                tasks = d.get("tasks") if isinstance(d, dict) else None
-                for t in (tasks or [])[:limit]:
-                    if isinstance(t, dict) and t.get("status") in ("wait", "doing"):
-                        lines.append("- [任务#%s] %s（%s）" % (
-                            t.get("id"), t.get("name") or "", t.get("status") or ""))
-            except Exception:
-                pass
-            try:
-                for b in list_bugs(cfg, profile.get("product"), profile)[:limit]:
-                    lines.append("- [Bug#%s] %s（严重级 %s）" % (
-                        b.get("id"), b.get("title") or "",
-                        b.get("severity") or b.get("status") or ""))
-            except Exception:
-                pass
-    except Exception:
-        return ""
-    return "\n".join(lines[:limit * 2])
-
-
-def fetch_modules(product_id):
-    """拉产品模块清单（模块路由配置辅助）。接口不存在/失败报人话，提示手填 ID。"""
-    _ensure_loaded()
-    cfg = _cfg()
-    try:
-        d = _call("GET", "/products/%s/modules" % product_id, cfg=cfg)
-    except ZenError as e:
-        msg = str(e)
-        # 只有「接口不存在」类错误才补「没有该接口」提示；老通道自己报的
-        # 空树/脚本回包已带人话结论，别再叠一层误导
-        if "404" in msg or "不认识" in msg:
-            msg += "——也可能你的禅道没有该接口：请在禅道产品视图 URL 里查模块 ID 手工填写"
-        return {"ok": False, "error": msg}
-    # 形状兼容器（{_list} / {modules:[...]} / {id:{...}} 字典 / 裸数组）
-    items = _list_items(d, "modules")
-    if items is None:
-        items = d.get("_list") if isinstance(d.get("_list"), list) else None
-    out = []
-    for m in items or []:
-        if isinstance(m, dict) and m.get("id"):
-            out.append({"id": m.get("id"), "name": str(m.get("name") or "")})
-    if not out:
-        if isinstance(d, dict) and set(d.keys()) == {"_list"}:
-            # 自家归一包装（裸数组进 _api/_old_route 时打的包）：空清单=
-            # 产品真没模块，不是形状问题，别拿顶层键吓人
-            return {"ok": False,
-                    "error": "禅道回了空模块清单（该产品可能没建模块）——请手工填模块 ID"}
-        # 带响应形状摘要帮排查（顶层键名，不吐正文——用户实测反馈
-        # 「响应形状不认识」却看不到真实形状，没法报修）
-        keys = ",".join(sorted(str(k) for k in d.keys())) if isinstance(d, dict) else type(d).__name__
-        return {"ok": False,
-                "error": "模块清单为空或响应形状不认识（响应顶层键：%s）——请手工填模块 ID" % (keys or "非字典")}
-    return {"ok": True, "modules": out[:200]}
-
-
-def _list_items(d, key):
-    """禅道 REST v1 清单响应形状兼容：{key:[...]} / 裸数组(_list) / {id:{...}} 字典。"""
-    if not isinstance(d, dict):
-        return None
-    if isinstance(d.get(key), list):
-        return d[key]
-    if isinstance(d.get("_list"), list):
-        return d["_list"]
-    vals = [v for v in d.values() if isinstance(v, dict) and v.get("id")]
-    return vals or None
-
-
-def fetch_products():
-    """拉产品清单（产品 ID 下拉选择辅助）。接口不存在/失败返回 ok:False+人话。"""
-    _ensure_loaded()
-    cfg = _cfg()
-    out, total = [], 0
-    try:
-        page = 1
-        while page <= 3:
-            d = _call("GET", "/products?page=%d&limit=100" % page, cfg=cfg)
-            items = _list_items(d, "products")
-            if not items:
-                break
-            out.extend(it for it in items
-                       if isinstance(it, dict) and str(it.get("id") or "") != "")
-            try:
-                total = int(d.get("total") or 0)
-            except (TypeError, ValueError, AttributeError):
-                total = 0
-            if not total or len(out) >= total:
-                break
-            page += 1
-    except ZenError as e:
-        return {"ok": False, "error": "%s——请先「测试连接」确认地址与账号可用" % e}
-    if not out:
-        return {"ok": False, "error": "禅道里没有产品，或响应形状不认识（REST API 需 ≥15）"}
-    seen, uniq = set(), []
-    for m in out:
-        try:
-            pid = int(m["id"])
-        except (TypeError, ValueError):
-            pid = m["id"]
-        if pid in seen:
-            continue
-        seen.add(pid)
-        uniq.append({"id": pid, "name": str(m.get("name") or ""),
-                     "status": str(m.get("status") or "")})
-    uniq.sort(key=lambda x: str(x["id"]))
-    return {"ok": True, "products": uniq[:500]}
-
-
-def fetch_users():
-    """拉禅道账号清单（负责人/转派下拉辅助）。失败返回 ok:False，前端仍可手填。"""
-    _ensure_loaded()
-    cfg = _cfg()
-    try:
-        d = _call("GET", "/users?limit=1000", cfg=cfg)
-    except ZenError as e:
-        return {"ok": False, "error": "%s——账号清单拉不到，仍可手填账号" % e}
-    items = None
-    if isinstance(d, dict):
-        if isinstance(d.get("users"), list):
-            items = d["users"]
-        elif isinstance(d.get("_list"), list):
-            items = d["_list"]
-        else:
-            items = []
-            for k, v in d.items():      # {账号: 用户} 字典形状
-                if not isinstance(v, dict):
-                    continue
-                u = dict(v)
-                u.setdefault("account", str(k))
-                items.append(u)
-    out, seen = [], set()
-    for u in items or []:
-        if not isinstance(u, dict):
-            continue
-        acct = str(u.get("account") or "").strip()
-        if not acct or acct in seen:
-            continue
-        seen.add(acct)
-        out.append({"account": acct, "realname": str(u.get("realname") or "")})
-    if not out:
-        return {"ok": False, "error": "账号清单为空或响应形状不认识——仍可手填账号"}
-    out.sort(key=lambda x: x["account"])
-    return {"ok": True, "users": out[:500]}
-
-
-def test_connection(base_url=None, account=None, password=None):
-    """连接测试：探明通道登录 + 拉第一个产品的 bug 列表（有档案时）。返回 (ok, 人话结果)。"""
-    c = _cfg()
-    base_url = str(base_url if base_url is not None else c.get("base_url") or "").strip()
-    account = str(account if account is not None else c.get("account") or "").strip()
-    password = str(password if password is not None else c.get("password") or "").strip()
-    if not (base_url and account and password):
-        return False, "地址、账号、密码都要填全"
-    b = base_url.strip().rstrip("/")
-    probe = {"base_url": b, "account": account, "password": password}
-    try:
-        tok = _rest_login(b, account, password)
-    except ZenError as e:
-        return False, str(e)
-    if tok:
-        _MODE[b] = "rest"
-    else:
-        try:
-            _old_login(probe, _REST_LAST_ERR.get("msg", ""))
-            _MODE[b] = "old"
-        except ZenError as e:
-            return False, str(e)
-    tag = "老版 JSON 接口" if _MODE.get(b) == "old" else "REST v1"
-    try:
-        profiles = _profiles(c)
-        if profiles:
-            bugs = list_bugs(probe, profiles[0]["product"], profiles[0])
-            return True, "连接成功（%s），产品 %s 可访问（当前 %d 条 bug 在列表里）" % (
-                tag, profiles[0]["product"], len(bugs))
-    except ZenError as e:
-        return False, "登录成功（%s），但拉 bug 列表失败：%s" % (tag, e)
-    return True, "连接成功（%s；未配产品档案，跳过列表探测）" % tag
-
-
-# ---------------------------------------------------------------- 排查（triage）
-
-def _ai_triage(bug, profile):
-    """AI 兜底排查：单次 LLM 调用判端。不可用/解析失败返回 None（bookmeta 同款配方）。
-
-    图片守卫（#27754 案）：描述指向截图而图片读不到 → unknown 留人工，绝不拿
-    纯文本盲判。排查模型不会看图时也不甩人工——自动全库找看图候选（声明
-    image_in 的优先，名字推断兜底），候选逐个试、实测吃图成功回写能力标记。
-    """
-    try:
-        from . import modelhub, runner
-        images = []
-        if _bug_mentions_images(bug):
-            try:
-                images = _bug_images(bug.get("id"), bug.get("steps"))
-            except Exception:
-                log.debug("zentao: bug 图片获取失败", exc_info=True)
-            if not images:
-                return {"side": "unknown",
-                        "reason": "描述指向截图/图片但图片未能读取，证据不足，"
-                                  "留人工确认后定责",
-                        "imgs": 0}
-        orch = modelhub.resolve_orchestrator()
-        prov, model = orch if orch else (None, None)
-        repos = profile.get("repos") or {}
-        hints = profile.get("repo_hints") or {}
-
-        def _repo_desc(side):
-            hint = str(hints.get(side) or "").strip()
-            if hint:
-                return hint
-            wd = str((repos.get(side) or {}).get("workdir") or "").strip()
-            if wd:
-                return "目录 " + Path(wd).name
-            return "（未配置）"
-
-        lines = ["你是缺陷分诊员：根据缺陷描述判断问题属于哪个仓库端。", "",
-                 "【缺陷】#%s %s" % (bug.get("id"), str(bug.get("title") or ""))]
-        steps = _strip_html(bug.get("steps"))
-        if steps:
-            lines.append(steps[:3000])
-        if images:
-            lines.append("【截图】随本消息附 %d 张 bug 截图，判定前先看图，"
-                         "截图与文字冲突时以截图为准。" % len(images))
-        lines.append("")
-        lines.append("【仓库背景】后端仓库：%s；前端仓库：%s" % (_repo_desc("backend"),
-                                                        _repo_desc("frontend")))
-        lines.append("")
-        lines.append('只输出 JSON（不要别的文字）：{"side": "backend|frontend|both|not_ours", '
-                     '"reason": "一句话依据"}')
-        lines.append("判定口径：backend=纯后端问题；frontend=纯前端问题；both=两端都要改；"
-                     "not_ours=与这两个仓库无关（第三方服务/环境/需求变更/数据问题等）。")
-        prompt = "\n".join(lines)
-        if images:
-            cand = modelhub.vision_candidates(
-                prefer_provider_id=(prov or {}).get("id") or "",
-                prefer_model=model or "")
-            if not cand:
-                return {"side": "unknown",
-                        "reason": "所有已启用模型都不支持图片输入，无法看图定责"
-                                  "（请在模型管理里启用一个视觉模型）",
-                        "imgs": 0}
-            res, last_err = None, ""
-            for p, m, inferred in cand:
-                r = modelhub.chat(p["id"], m, prompt, max_tokens=500, timeout=90,
-                                  images=images)
-                if r.get("ok"):
-                    res = r
-                    if inferred:
-                        # 名字推断的候选实测吃图成功：回写能力标记（幂等），
-                        # 模型管理页图徽章同步可见，后续调用直接命中
-                        err = modelhub.set_model_caps(p["id"], m, True)
-                        if err:
-                            log.debug("zentao: 自动标记 image_in 失败：%s", err)
-                    break
-                last_err = str(r.get("error") or "")
-                log.warning("zentao: 视觉排查候选 %s/%s 不可用：%s",
-                            p.get("id"), m, last_err[:160])
-            if not res:
-                return {"side": "unknown",
-                        "reason": "看图排查全部候选失败（%s）" % last_err[:160],
-                        "imgs": 0}
-        else:
-            if not orch:
-                return None
-            res = modelhub.chat(prov["id"], model, prompt, max_tokens=500, timeout=90)
-            if not res.get("ok"):
-                log.warning("zentao: AI 排查失败：%s", res.get("error"))
-                return None
-        data = runner.extract_json(res.get("text") or "")
-        side = str((data or {}).get("side") or "").strip().lower()
-        if side not in TRIAGE_SIDES:
-            return None
-        return {"side": side, "reason": str((data or {}).get("reason") or "")[:300],
-                "imgs": len(images)}
-    except Exception:
-        log.debug("zentao: AI 排查异常", exc_info=True)
-        return None
-
-
-def _triage(bug, profile, cfg):
-    """排查定责：模块路由 > AI > unknown。返回 {side, reason, by, account}。"""
-    mid = str(bug.get("module") or "")
-    if mid:
-        for r in profile.get("module_routes") or []:
-            try:
-                if int(r.get("module") or 0) != int(mid):
-                    continue
-            except (TypeError, ValueError):
-                continue
-            side = str(r.get("side") or "")
-            if side in TRIAGE_SIDES:
-                return {"side": side, "reason": "模块 #%s 路由规则" % mid,
-                        "by": "rule", "account": str(r.get("account") or "").strip()}
-    if cfg.get("triage_ai"):
-        res = _ai_triage(bug, profile)
-        if res:
-            res["by"] = "ai"
-            res["account"] = ""
-            return res
-    return {"side": "unknown", "reason": "模块未命中路由且 AI 排查不可用", "by": "fallback",
-            "account": ""}
-
-
-# ---------------------------------------------------------------- 修复任务拉起
-
-def _repo_of(profile, side):
-    return (profile.get("repos") or {}).get(side) or dict(_REPO_DEFAULTS)
-
-
-def _goal_text(bug, side=None, images=0):
-    """bug → 修复目标提示词。side 给出时附端约束；images>0 时提示先读附件截图。"""
-    bid = bug.get("id")
-    lines = ["修复禅道 Bug #%s：%s" % (bid, str(bug.get("title") or "").strip())]
-    steps = _strip_html(bug.get("steps"))
-    if steps:
-        lines.append("")
-        lines.append("【重现步骤/问题描述】")
-        lines.append(steps[:4000])
-    sev, pri = _severity(bug), str(bug.get("pri") or "").strip()
-    meta = []
-    if sev:
-        meta.append("严重度 %s（1 最严重）" % sev)
-    if pri:
-        meta.append("优先级 %s" % pri)
-    if str(bug.get("module") or "").strip():
-        meta.append("模块 #%s" % bug["module"])
-    env = " / ".join(x for x in (str(bug.get("os") or "").strip(),
-                                 str(bug.get("browser") or "").strip()) if x)
-    if env:
-        meta.append("环境 " + env)
-    kw = str(bug.get("keywords") or "").strip()
-    if kw:
-        meta.append("关键字 " + kw)
-    if meta:
-        lines.append("")
-        lines.append("【元信息】" + "；".join(meta))
-    if images:
-        lines.append("")
-        lines.append("【截图】bug 截图共 %d 张，已放在工作目录 _attachments/（文件名 "
-                     "zenbug*）；先读图再动手——需求以截图为准，不要凭文字想象。"
-                     % images)
-    lines.append("")
-    lines.append("【要求】只修这个 bug，不做无关重构；改动最小化；修完自查不引入回归。")
-    if side in SIDES:
-        lines.append("【端约束】本任务只负责【%s】部分；%s部分由别人处理，不要越界改动。"
-                     % (SIDE_CN[side],
-                        SIDE_CN["frontend" if side == "backend" else "backend"]))
-    lines.append("完成后给出修改说明。")
-    return "\n".join(lines)
-
-
-def _launch_fix(bug, profile, side, cfg):
-    """为一个 bug 的某一端建修复任务并立即启动。"""
-    bid = bug.get("id")
-    repo = _repo_of(profile, side)
-    wd = str(repo.get("workdir") or "").strip() or settings.default_workdir()
-    title = ("[禅道#%s][%s] %s" % (bid, SIDE_CN.get(side, side),
-                                   str(bug.get("title") or "").strip())).strip()[:60]
-    # bug 截图随任务走：修复智能体跟排查同样需要看图，否则修出来的就是想象中的 bug
-    imgs = []
-    try:
-        if _bug_mentions_images(bug):
-            imgs = _bug_images(bid, bug.get("steps"))
-    except Exception:
-        imgs = []
-    atts = _save_bug_attachments(str(wd), bid, imgs)
-    payload = {"type": "code", "title": title, "goal": _goal_text(bug, side, images=len(atts)),
-               "workdir": wd}
-    if atts:
-        payload["attachments"] = atts
-    base = str(repo.get("git_rev") or "").strip()
-    if not base:
-        # 档案没配基线就取落单时点的 HEAD：让修复走任务分支隔离——修完自动
-        # 提交在任务分支上，对账合并才真正落库，resolve 才站得住（#27697 案：
-        # 无基线任务修完只躺在工作区，闭环永远不会替你提交）。
-        base = _head_rev(wd)
-    if base:
-        payload["git_rev"] = base
-    if str(repo.get("verify_command") or "").strip():
-        payload["verify_command"] = str(repo["verify_command"]).strip()
-    task = None
-    run = None
-    try:
-        task = store.create_task(payload)
-        run = store.create_run("orchestration", task["title"], task_id=task["id"])
-        store.update_task_status(task["id"], "queued")
-        jobs.enqueue({"kind": "orchestration", "run_id": run["id"],
-                      "task_id": task["id"]})
-        return task, run
-    except Exception:
-        log.exception("zentao: 修复任务启动失败 bug=%s side=%s task=%s run=%s",
-                      bid, side, (task or {}).get("id"), (run or {}).get("id"))
-        if run:
-            try:
-                store.update_run(run["id"], status="failed",
-                                 error="禅道修复任务启动失败，本次未排队，请稍后重试",
-                                 ended_at=time.strftime("%Y-%m-%d %H:%M:%S"))
-            except Exception:
-                log.exception("zentao: 修复运行失败收口失败 run=%s", run.get("id"))
-        if task:
-            try:
-                store.update_task_status(task["id"], "failed")
-            except Exception:
-                log.exception("zentao: 修复任务失败收口失败 task=%s", task.get("id"))
-        raise
-
-
-# ---------------------------------------------------------------- 回写文本
-
-def _diffstat(workdir, task_id):
-    """任务分支相对基线的改动统计（人话一行）。拿不到返回空串。"""
-    try:
-        from . import gitmod, runner
-        if task_id and workdir:
-            br = gitmod.branch_name(task_id)
-            r = runner.run_process(
-                argv=["git", "-C", workdir, "diff", "--shortstat", "HEAD..." + br],
-                timeout=20)
-            if r.get("ok"):
-                return (r.get("stdout") or "").strip()
-    except Exception:
-        log.debug("zentao: diffstat 失败", exc_info=True)
-    return ""
-
-
-def _fix_summary(claim, run, profile, side):
-    """我方某一端的修复摘要（转派评论与 resolve 报告共用的事实部分）。"""
-    lines = []
-    git = (run or {}).get("git") or {}
-    commit = str(git.get("commit") or "").strip()
-    from_branch = str(git.get("from_branch") or "").strip()
-    tid = _task_of(claim, side).get("task_id") or ""
-    if commit:
-        lines.append("【%s】修复提交 %s%s" % (SIDE_CN.get(side, side), commit,
-                     "（已合并回 %s）" % from_branch if from_branch else ""))
-    stat = _diffstat(str(_repo_of(profile, side).get("workdir") or ""), tid)
-    if stat:
-        lines.append("【%s】改动统计：%s" % (SIDE_CN.get(side, side), stat))
-    v = (run or {}).get("verdict") or {}
-    if v:
-        lines.append("【%s】验证结论：%s" % (SIDE_CN.get(side, side),
-                     "通过（评审达标）" if v.get("pass") or v.get("publishable") else "完成"))
-    return lines
-
-
-def _task_of(claim, side):
-    for t in claim.get("tasks") or []:
-        if t.get("side") == side:
-            return t
-    return {}
-
-
-def _report_text(claim, runs, profile, cfg):
-    """resolve 评论：确定性事实（逐端汇总）。"""
-    tri = claim.get("triage") or {}
-    lines = ["【CodeBee 自动修复报告】",
-             "Bug：#%s %s" % (claim.get("bug_id"), claim.get("title") or "")]
-    if tri.get("side"):
-        lines.append("排查结论：%s问题（%s）" % (tri.get("side"),
-                                               tri.get("reason") or tri.get("by") or "按规则"))
-    for t in claim.get("tasks") or []:
-        lines.extend(_fix_summary(claim, runs.get(t.get("run_id")), profile, t.get("side")))
-    lines.append("（本条由 CodeBee 禅道集成自动回写）")
-    return "\n".join(lines)
-
-
-def _transfer_text(claim, profile, fixed_runs, target_side):
-    """转派评论：排查结论 + 我方已做工作 + 请对方继续。"""
-    tri = claim.get("triage") or {}
-    lines = ["【CodeBee 排查转派】",
-             "Bug：#%s %s" % (claim.get("bug_id"), claim.get("title") or ""),
-             "排查结论：%s问题——%s" % (SIDE_CN.get(target_side, target_side),
-                                      tri.get("reason") or tri.get("by") or "按规则")]
-    if tri.get("imgs"):
-        lines.append("（判定依据含 %d 张 bug 截图）" % tri["imgs"])
-    for side, run in (fixed_runs or []):
-        lines.extend(_fix_summary(claim, run, profile, side))
-    lines.append("请%s负责人接手处理；本 bug 保持激活，处理完请按正常流程解决。"
-                 % SIDE_CN.get(target_side, target_side))
-    lines.append("（本条由 CodeBee 禅道集成自动回写）")
-    return "\n".join(lines)
-
-
-def _fail_text(claim, failed_tasks, runs, internal=False):
-    tri = claim.get("triage") or {}
-    lines = ["【CodeBee 自动修复未成功】",
-             "Bug：#%s %s" % (claim.get("bug_id"), claim.get("title") or "")]
-    if tri.get("side"):
-        lines.append("排查结论：%s（%s）" % (tri.get("side"), tri.get("reason") or ""))
-    for t in failed_tasks:
-        r = runs.get(t.get("run_id")) or {}
-        why = str(r.get("error") or "").strip() or ("运行状态 " + str(r.get("status") or ""))
-        lines.append("【%s】失败：%s" % (SIDE_CN.get(t.get("side"), t.get("side")), why[:300]))
-    if internal:
-        lines.append("（失败原因是 CodeBee 工具自身异常，不代表修复结论；"
-                     "请人工排查 CodeBee 或续跑修复任务）")
-    lines.append("CodeBee 修复任务：%s（可人工续跑或接管）；本 bug 保持待处理。"
-                 % "、".join(t.get("task_id") or "?" for t in claim.get("tasks") or []))
-    return "\n".join(lines)
-
-
-# ---------------------------------------------------------------- 禅道写回动作
-
-def _bug_opened_by(cfg, bug_id):
-    try:
-        d = _call("GET", "/bugs/%s" % bug_id, cfg=cfg)
-        return _acct(d.get("openedBy"))
-    except ZenError:
-        return ""
-
-
-def _ensure_resolved(cfg, bug_id, comment, assign_to):
-    """resolve（幂等）：已是 resolved/closed 视为成功。"""
-    cur = _call("GET", "/bugs/%s" % bug_id, cfg=cfg)
-    if str(cur.get("status") or "") in ("resolved", "closed"):
-        return True
-    body = {"resolution": "fixed", "resolvedBuild": "trunk", "comment": comment}
-    if assign_to:
-        body["assignedTo"] = assign_to
-    operation_id = operations.begin("zentao:bug:%s:resolve" % bug_id, body,
-                                    metadata={"bug_id": bug_id, "action": "resolve"})
-    try:
-        result = _call("POST", "/bugs/%s/resolve" % bug_id, cfg=cfg, body=body)
-        receipt = str((result or {}).get("id") or bug_id)
-        operations.confirm(operation_id, remote_receipt=receipt,
-                           metadata={"bug_id": bug_id, "action": "resolve"})
-    except Exception as exc:
-        operations.finish_exception(operation_id, exc)
-        raise
-    return True
-
-
-def _transfer(cfg, bug_id, target, comment):
-    """转派：PUT assignedTo+comment（bug 保持激活）。target 空=只评论。抛 ZenError。
-
-    禅道 PUT /bugs/{id} 不带 assignedTo 会把指派人清空（#27783 案：只评论
-    也出了「指派给(空)」记录），空 target 必须先 GET 带回当前指派人；GET
-    不到就整个不发——宁可不评论，不可盲写清指派。
-    """
-    body = {"comment": comment}
-    if target:
-        body["assignedTo"] = target
-    else:
-        cur = _call("GET", "/bugs/%s" % bug_id, cfg=cfg)
-        body["assignedTo"] = _acct(cur.get("assignedTo")) or ""
-    operation_id = operations.begin("zentao:bug:%s:transfer" % bug_id, body,
-                                    metadata={"bug_id": bug_id, "action": "transfer"})
-    try:
-        result = _call("PUT", "/bugs/%s" % bug_id, cfg=cfg, body=body)
-        receipt = str((result or {}).get("id") or bug_id)
-        operations.confirm(operation_id, remote_receipt=receipt,
-                           metadata={"bug_id": bug_id, "action": "transfer"})
-    except Exception as exc:
-        operations.finish_exception(operation_id, exc)
-        raise
-
-
-def _notify(text):
-    try:
-        from . import notify
-        notify.push_text(text)
-    except Exception:
-        log.debug("zentao: 群通知失败", exc_info=True)
-
-
-def _merge_branch(workdir, task):
-    try:
-        from . import gitmod
-        return gitmod.merge_task_branch(workdir, task)
-    except Exception as e:
-        return False, "合并异常：%s" % e, None
-
-
-def _head_rev(workdir):
-    """工作目录当前 HEAD 短哈希（作任务隔离基线）。非仓库/拿不到返回空串。"""
-    try:
-        from . import gitmod, runner
-        r = runner.run_process(
-            argv=["git", "-C", str(workdir), "rev-parse", "--short", "HEAD"], timeout=20)
-        if r.get("ok"):
-            rev = (r.get("stdout") or "").strip()
-            if rev and gitmod.valid_rev(rev):
-                return rev
-    except Exception:
-        log.debug("zentao: 读取 HEAD 失败", exc_info=True)
-    return ""
-
-
-def _uncommitted(workdir):
-    """工作目录里有没有未提交的已跟踪改动（untracked 不算——CLI 噪音文件太多）。
-
-    返回 (是否脏, 人话明细)。目录不是 git 仓库视为不脏（无提交语义，放行）。
-    """
-    try:
-        from . import runner
-        if not workdir:
-            return False, ""
-        r = runner.run_process(
-            argv=["git", "-C", str(workdir), "status", "--porcelain", "-uno", "--", "."],
-            timeout=20)
-        if r.get("ok"):
-            dirty = [ln.strip() for ln in (r.get("stdout") or "").splitlines() if ln.strip()]
-            if dirty:
-                return True, "未提交改动 %d 处（如 %s）" % (len(dirty), dirty[0][:80])
-    except Exception:
-        log.debug("zentao: 工作区状态检查失败", exc_info=True)
-    return False, ""
-
-
-def _set_claim(bid, **patch):
-    with _LOCK:
-        c = _STATE["claims"].get(str(bid))
-        if c is None:
-            return
-        c.update(patch)
-        c["at"] = time.strftime("%Y-%m-%d %H:%M:%S")
-        _save_locked()
-
-
-def _route_account(profile, tri, side):
-    """转派/升级目标：模块路由 account > 端负责人。"""
-    if tri and tri.get("account"):
-        return str(tri["account"]).strip()
-    return str((profile.get("owners") or {}).get(side) or "").strip()
-
-
-# ---------------------------------------------------------------- 对账回写
-
-def _finish_ok(claim, cfg):
-    """我方任务全部 done：合并 → （需要则）转派 → 否则 resolve。"""
-    bid = str(claim.get("bug_id"))
-    profile = _profile_for(cfg, claim.get("product")) or {}
-    tri = claim.get("triage") or {}
-    tasks = claim.get("tasks") or []
-    runs = {t.get("run_id"): store.get_run(t.get("run_id") or "") for t in tasks}
-    # 1) 逐任务合并（任一失败即停：不转派不 resolve，绝不带着没落库的修复转派）
-    if not cfg.get("auto_merge"):
-        _set_claim(bid, state="done_manual",
-                   note="修复完成；auto_merge 已关，请人工合并落库后到禅道解决 bug")
-        return
-    for t in tasks:
-        task = store.get_task(t.get("task_id") or "")
-        if task is None or not task.get("git_rev"):
-            continue
-        ok, err, _info = _merge_branch(str(_repo_of(profile, t.get("side")).get("workdir")
-                                           or settings.default_workdir()), task)
-        if not ok:
-            _set_claim(bid, state="merge_failed",
-                       note="【%s】任务分支合并失败：%s（不转派不 resolve，留人工）"
-                            % (SIDE_CN.get(t.get("side"), t.get("side")), err))
-            _notify("🐛❌ 禅道 Bug #%s 的%s修复代码合并失败：%s\n修复任务：%s"
-                    % (bid, SIDE_CN.get(t.get("side"), ""), err, t.get("task_id") or "?"))
-            return
-    # 1.5) 落库闸：没基线的任务不走任务分支隔离，改动只躺在工作区——带着
-    # 未提交的已跟踪改动点 resolve 是对禅道撒谎（#27697 案），拦下转人工。
-    for t in tasks:
-        task = store.get_task(t.get("task_id") or "")
-        if task is None or task.get("git_rev"):
-            continue
-        dirty, detail = _uncommitted(str(_repo_of(profile, t.get("side")).get("workdir")
-                                         or settings.default_workdir()))
-        if dirty:
-            _set_claim(bid, state="done_manual",
-                       note="修复完成但工作区有未提交改动（%s），已阻止自动 resolve"
-                            "——请人工提交后到禅道解决" % detail)
-            _notify("🐛⚠️ 禅道 Bug #%s 修复完成但未提交落库，已阻止自动 resolve\n%s"
-                    % (bid, claim.get("title") or ""))
-            return
-    # 2) 需要转派的端 = 判定端 - 我方端（both 且我方只管一端时非空）
-    verdict_sides = ([tri["side"]] if tri.get("side") in SIDES
-                     else list(SIDES) if tri.get("side") == "both" else [])
-    other_sides = [s for s in verdict_sides if s not in (profile.get("our_sides") or [])]
-    if other_sides:
-        targets = {}
-        for s in other_sides:
-            tgt = _route_account(profile, tri, s)
-            if not tgt:
-                _set_claim(bid, state="need_manual",
-                           note="我方已修完，但未配置%s负责人，无法转派——请人工转派" % SIDE_CN[s])
-                _notify("🐛⚠️ 禅道 Bug #%s 我方部分已修完，但未配置%s负责人，请人工转派"
-                        % (bid, SIDE_CN[s]))
-                return
-            targets[s] = tgt
-        fixed_runs = [(t.get("side"), runs.get(t.get("run_id"))) for t in tasks]
-        for s in other_sides:
-            try:
-                _transfer(cfg, bid, targets[s], _transfer_text(claim, profile, fixed_runs, s))
-            except ZenError as e:
-                attempts = int(claim.get("attempts") or 0) + 1
-                if attempts >= RESOLVE_MAX_ATTEMPTS:
-                    _set_claim(bid, attempts=attempts, state="resolve_failed",
-                               note="转派连续 %d 次失败：%s" % (attempts, e))
-                    _notify("🐛⚠️ 禅道 Bug #%s 修完但转派失败：%s" % (bid, e))
-                else:
-                    _set_claim(bid, attempts=attempts, state="fixing",
-                               note="转派第 %d 次失败，下轮重试：%s" % (attempts, e))
-                return
-        _set_claim(bid, state="transferred",
-                   note="我方（%s）已修完并合并，转派给 %s"
-                        % ("/".join(SIDE_CN.get(t.get("side"), "") for t in tasks),
-                           "、".join(SIDE_CN[s] + ":" + targets[s] for s in other_sides)))
-        _notify("🐛🔁 禅道 Bug #%s 我方已修完，转派 %s\n%s"
-                % (bid, "、".join(SIDE_CN[s] + ":" + targets[s] for s in other_sides),
-                   claim.get("title") or ""))
-        return
-    # 3) 全部我方端：resolve（幂等）+ 指回报告人
-    if not cfg.get("auto_resolve"):
-        _set_claim(bid, state="done_manual",
-                   note="修复完成；auto_resolve 已关，请人工确认后到禅道解决 bug")
-        return
-    report = _report_text(claim, runs, profile, cfg)
-    opened = _bug_opened_by(cfg, bid)
-    try:
-        _ensure_resolved(cfg, bid, report, opened)
-    except ZenError as e:
-        attempts = int(claim.get("attempts") or 0) + 1
-        if attempts >= RESOLVE_MAX_ATTEMPTS:
-            _set_claim(bid, attempts=attempts, state="resolve_failed",
-                       note="resolve 连续 %d 次失败：%s" % (attempts, e))
-            _notify("🐛⚠️ 禅道 Bug #%s 修复完成但回写禅道失败：%s" % (bid, e))
-        else:
-            _set_claim(bid, attempts=attempts, state="fixing",
-                       note="resolve 第 %d 次失败，下轮重试：%s" % (attempts, e))
-        return
-    _set_claim(bid, state="resolved", note="已 resolve(fixed)%s"
-               % ("，指回报告人 " + opened if opened else ""))
-    _notify("🐛✅ 禅道 Bug #%s 已修复并 resolve\n%s\n修复任务：%s"
-            % (bid, claim.get("title") or "", "、".join(t.get("task_id") or "" for t in tasks)))
-
-
-# CodeBee 自身崩溃形态：run.error 是裸 Python 异常 repr（如 NameError(...)，
-# 0.1.63 _TRANSIENT 缩进事故实测形态）或带 Python traceback。这不是「修不动
-# bug」——转派给端负责人等于拿自家工具故障甩锅。冒号形态（TimeoutError: x）
-# 多为 CLI/被测仓的真实输出，不算内部崩溃。
-_CRASH_REPR_RE = re.compile(r"^[A-Za-z_]\w*(?:Error|Exception|Interrupt)\(")
-
-
-def _looks_internal_crash(err):
-    e = str(err or "").strip()
-    if not e:
-        return False
-    return ("Traceback (most recent call last" in e
-            or _CRASH_REPR_RE.match(e) is not None)
-
-
-def _finish_failed(claim, cfg):
-    """我方任一任务失败：评论尝试记录 + 转派该端负责人（有配则转）。
-
-    CodeBee 自身崩溃（内部异常）不转派——工具故障不是「修不动」，转出去是
-    甩锅，只评论+群通知人工。负责人未配置同样只评论。
-    """
-    bid = str(claim.get("bug_id"))
-    profile = _profile_for(cfg, claim.get("product")) or {}
-    tasks = claim.get("tasks") or []
-    runs = {t.get("run_id"): store.get_run(t.get("run_id") or "") for t in tasks}
-    failed = [t for t in tasks
-              if str((runs.get(t.get("run_id")) or {}).get("status") or "") not in
-              ("queued", "running", "done")]
-    internal = any(_looks_internal_crash((runs.get(t.get("run_id")) or {}).get("error"))
-                   for t in failed)
-    text = _fail_text(claim, failed, runs, internal=internal)
-    # 升级目标：取第一个失败端的路由账号/负责人（内部崩溃不转派）
-    target = ""
-    if not internal:
-        for t in failed:
-            target = _route_account(profile, claim.get("triage") or {}, t.get("side"))
-            if target:
-                break
-    try:
-        _transfer(cfg, bid, target, text)
-        if internal:
-            note = "CodeBee 内部异常，已评论说明（不转派，请人工排查工具链）"
-            state = "commented"
-        elif target:
-            note = "已评论说明并转派 %s" % target
-            state = "escalated"
-        else:
-            side = str((failed[0].get("side") if failed else "") or "")
-            note = "已评论说明（%s负责人未配置，未转派）" % SIDE_CN.get(side, side or "该端")
-            state = "commented"
-    except ZenError as e:
-        log.warning("zentao: bug %s 失败评论未送达（群通知兜底）：%s", bid, e)
-        note = "修复失败，评论未送达：%s" % e
-        state = "commented"
-    _set_claim(bid, state=state, note=note)
-    why = ("CodeBee 内部异常，未转派" if internal
-           else ("已转派 " + target if target else "负责人未配置，未转派"))
-    _notify("🐛❌ 禅道 Bug #%s 自动修复未成功（%s）\n%s\n修复任务：%s"
-            % (bid, why,
-               claim.get("title") or "", "、".join(t.get("task_id") or "" for t in tasks)))
-
-
-def _refresh_claim_runs(claim):
-    """把 claim.tasks 的 run_id 对齐到各任务最新一次运行，返回 (tasks, 是否换绑)。
-
-    领单时钉住的 run 失败后，人工重试会换新 run——对账只认旧 run 会把已经
-    成功的修复判成失败（2026-09-22 #27697 误报失败评论案）。
-    """
-    fresh, changed = [], False
-    for t in claim.get("tasks") or []:
-        tid = str(t.get("task_id") or "")
-        if tid:
-            runs = store.task_runs(tid)
-            rid = str((runs[0].get("id") if runs else "") or "")
-            if rid and rid != t.get("run_id"):
-                t = dict(t, run_id=rid)
-                changed = True
-        fresh.append(t)
-    return fresh, changed
-
-
-def _reconcile(cfg):
-    """对账：fixing 中的 claim 查各 run 终态并回写。单条异常只跳过该条。"""
-    with _LOCK:
-        fixing = [dict(c) for c in _STATE["claims"].values() if c.get("state") == "fixing"]
-    for claim in fixing:
-        bid = str(claim.get("bug_id"))
-        try:
-            tasks, moved = _refresh_claim_runs(claim)
-            if moved:
-                _set_claim(bid, tasks=tasks)
-                claim = dict(claim, tasks=tasks)
-            runs = {t.get("run_id"): store.get_run(t.get("run_id") or "") for t in tasks}
-            if any(r is None for r in runs.values()):
-                _set_claim(bid, state="lost", note="运行记录不存在（可能被清理）")
-                continue
-            states = [str((r or {}).get("status") or "") for r in runs.values()]
-            if any(s in ("queued", "running") for s in states):
-                continue
-            if all(s == "done" for s in states):
-                _finish_ok(claim, cfg)
-            else:
-                _finish_failed(claim, cfg)
-        except Exception:
-            log.exception("zentao: claim %s 对账异常，跳过", bid)
-
-
-# ---------------------------------------------------------------- 扫描主流程
-
-def _sides_to_fix(profile, tri):
-    """判定端 ∩ 我方端。"""
-    tri_side = tri.get("side")
-    verdict_sides = [tri_side] if tri_side in SIDES else (list(SIDES) if tri_side == "both" else [])
-    return [s for s in verdict_sides if s in (profile.get("our_sides") or [])]
-
-
-def _route_one(bug, profile, cfg, notify=True):
-    """认领一个 bug：排查 → 分流（建任务/转派/留人工）。notify=False 用于
-    need_manual 存量的重排查（仍无解时不重复群通知）。返回动作文案或 None。"""
-    bid = str(bug.get("id") or "")
-    tri = _triage(bug, profile, cfg)
-    base_claim = {
-        "bug_id": bug.get("id"), "product": profile.get("product"),
-        "title": str(bug.get("title") or "")[:120],
-        "triage": tri, "tasks": [], "state": "fixing", "note": "",
-        "attempts": 0, "claimed_at": time.strftime("%Y-%m-%d %H:%M:%S"),
-    }
-    opened = _acct(bug.get("openedBy"))
-
-    if tri["side"] == "unknown":
-        base_claim["state"] = "need_manual"
-        base_claim["note"] = tri.get("reason") or "排查失败，留人工"
-        with _LOCK:
-            _STATE["claims"][bid] = base_claim
-            _save_locked()
-        if notify:
-            _notify("🐛❓ 禅道 Bug #%s 排查不出归属端（%s），留人工\n%s"
-                    % (bid, tri.get("reason") or "", bug.get("title") or ""))
-        return "need_manual"
-
-    if tri["side"] == "not_ours":
-        target = tri.get("account") or str(profile.get("owners").get("not_ours") or "") \
-            or opened
-        text = ("【CodeBee 排查转派】\nBug：#%s %s\n排查结论：非我方两个仓库的问题——%s\n"
-                "%s转回 %s 核实处理。\n（本条由 CodeBee 禅道集成自动回写）"
-                % (bid, bug.get("title") or "", tri.get("reason") or "按规则",
-                   ("（判定依据含 %d 张 bug 截图）\n" % tri["imgs"]) if tri.get("imgs") else "",
-                   target or "报告人"))
-        try:
-            _transfer(cfg, bid, target, text)
-            note = "非我方，已转派 %s" % (target or "（未指派，仅评论）")
-        except ZenError as e:
-            base_claim.update({"state": "need_manual", "note": "非我方转派失败：%s" % e})
-            with _LOCK:
-                _STATE["claims"][bid] = base_claim
-                _save_locked()
-            _notify("🐛⚠️ 禅道 Bug #%s 判定非我方但转派失败：%s" % (bid, e))
-            return "need_manual"
-        base_claim.update({"state": "transferred", "note": note})
-        with _LOCK:
-            _STATE["claims"][bid] = base_claim
-            _save_locked()
-        _notify("🐛↩️ 禅道 Bug #%s 判定非我方，已转派 %s\n%s"
-                % (bid, target or "报告人", bug.get("title") or ""))
-        return "transferred"
-
-    sides = _sides_to_fix(profile, tri)
-    if not sides:
-        # 纯对方端问题（含测试指错到我方账号的）：直接转派正确负责人
-        other = tri["side"] if tri["side"] in SIDES else None
-        if other is None:            # both 但我方两端都没配：整单转不了，留人工
-            base_claim["state"] = "need_manual"
-            base_claim["note"] = "双端问题但产品档案未配置我方端，请人工处理"
-            with _LOCK:
-                _STATE["claims"][bid] = base_claim
-                _save_locked()
-            _notify("🐛❓ 禅道 Bug #%s 为双端问题但未配我方端，留人工" % bid)
-            return "need_manual"
-        target = _route_account(profile, tri, other)
-        if not target:
-            base_claim["state"] = "need_manual"
-            base_claim["note"] = "判定为%s问题，但未配置%s负责人，无法转派" % (
-                SIDE_CN[other], SIDE_CN[other])
-            with _LOCK:
-                _STATE["claims"][bid] = base_claim
-                _save_locked()
-            _notify("🐛⚠️ 禅道 Bug #%s 判定为%s问题但未配负责人，请人工转派"
-                    % (bid, SIDE_CN[other]))
-            return "need_manual"
-        text = _transfer_text({"bug_id": bug.get("id"), "title": bug.get("title") or "",
-                               "triage": tri, "tasks": []}, profile, [], other)
-        try:
-            _transfer(cfg, bid, target, text)
-        except ZenError as e:
-            base_claim["state"] = "need_manual"
-            base_claim["note"] = "%s转派失败：%s" % (SIDE_CN[other], e)
-            with _LOCK:
-                _STATE["claims"][bid] = base_claim
-                _save_locked()
-            _notify("🐛⚠️ 禅道 Bug #%s 判定%s问题但转派失败：%s" % (bid, SIDE_CN[other], e))
-            return "need_manual"
-        base_claim.update({"state": "transferred",
-                           "note": "%s问题（测试原指派 %s），已转派 %s"
-                                   % (SIDE_CN[other], _acct(bug.get("assignedTo")) or "?", target)})
-        with _LOCK:
-            _STATE["claims"][bid] = base_claim
-            _save_locked()
-        _notify("🐛🔁 禅道 Bug #%s 判定%s问题，已转派 %s（原指派 %s）\n%s"
-                % (bid, SIDE_CN[other], target, _acct(bug.get("assignedTo")) or "?",
-                   bug.get("title") or ""))
-        return "transferred"
-
-    # 我方端：逐端建修复任务（任一端建不起来 → 整单留人工，不做半截修复）
-    tasks = []
-    try:
-        for side in sides:
-            task, run = _launch_fix(bug, profile, side, cfg)
-            tasks.append({"side": side, "task_id": task["id"], "run_id": run["id"],
-                          "state": "fixing"})
-    except Exception as e:
-        log.warning("zentao: bug %s 建修复任务失败：%s", bid, e)
-        base_claim["state"] = "need_manual"
-        base_claim["note"] = "建修复任务失败：%s" % e
-        with _LOCK:
-            _STATE["claims"][bid] = base_claim
-            _save_locked()
-        _notify("🐛⚠️ 禅道 Bug #%s 建修复任务失败：%s" % (bid, e))
-        return "need_manual"
-    base_claim["tasks"] = tasks
-    with _LOCK:
-        _STATE["claims"][bid] = base_claim
-        _save_locked()
-    _notify("🐛🔍 认领禅道 Bug #%s（%s问题·%s）：%s\n修复任务：%s"
-            % (bid, tri["side"], "由规则" if tri.get("by") == "rule" else "AI 判定",
-               bug.get("title") or "", "、".join(t["task_id"] for t in tasks)))
-    return "fixing"
-
-
-def _scan(cfg):
-    """按产品档案逐个拉 bug：认领新 bug + 重试 need_manual 的存量。返回认领数。"""
-    claimed = 0
-    profiles = _profiles(cfg)
-    if not profiles:
-        raise ZenError("未配置产品档案——请先在设置页添加产品并配置仓库")
-    for profile in profiles:
-        bugs = list_bugs(cfg, profile["product"], profile)
-        by_id = {str(b.get("id") or ""): b for b in bugs}
-        with _LOCK:
-            seen = set(_STATE["claims"].keys())
-            retry_ids = [k for k, c in _STATE["claims"].items()
-                         if c.get("state") == "need_manual" and k in by_id
-                         and str(by_id[k].get("status") or "") == "active"]
-        for bug in bugs:
-            bid = str(bug.get("id") or "")
-            if not bid or bid in seen or not _claimable(bug, profile):
-                continue
-            _route_one(bug, profile, cfg)
-            claimed += 1
-        # need_manual 重排查（bug 仍激活才出现在列表里）
-        for bid in retry_ids:
-            with _LOCK:
-                if _STATE["claims"].get(bid, {}).get("state") != "need_manual":
-                    continue
-            _route_one(by_id[bid], profile, cfg, notify=False)
-    return claimed
-
-
-def _poll(force=False):
-    """执行一次扫描；同一进程内只允许一个扫描实例。"""
-    if not _SCAN_LOCK.acquire(blocking=False):
-        return {"ok": False, "claimed": 0, "reconciled": False,
-                "error": "扫描正在进行，请稍候", "skipped": "in_progress"}
-    try:
-        return _poll_unlocked(force=force)
-    finally:
-        _SCAN_LOCK.release()
-
-
-def _poll_unlocked(force=False):
-    """一次完整轮询：对账回写 + （到点/强制时）扫描认领。异常不外抛。"""
-    _ensure_loaded()
-    with _LOCK:
-        cfg = _cfg()
-    out = {"ok": True, "claimed": 0, "reconciled": True, "error": ""}
-    try:
-        _reconcile(cfg)
-    except Exception:
-        log.exception("zentao: 对账异常")
-    if not force:
-        if not cfg.get("poll_enabled"):
-            return {"ok": True, "claimed": 0, "reconciled": True, "error": "",
-                    "skipped": "poll_disabled"}
-        with _LOCK:
-            nxt = _STATE.get("next_scan") or ""
-        if nxt:
-            try:
-                if datetime.fromisoformat(nxt) > datetime.now():
-                    return {"ok": True, "claimed": 0, "reconciled": True,
-                            "error": "", "skipped": "not_due"}
-            except ValueError:
-                pass
-    err = ""
-    try:
-        out["claimed"] = _scan(cfg)
-    except ZenError as e:
-        err = str(e)
-    except Exception as e:
-        err = "扫描失败：%s" % e
-        log.warning("zentao: %s", err)
-    now = datetime.now()
-    delay = timedelta(minutes=max(INTERVAL_MIN, int(cfg.get("interval_minutes") or INTERVAL_DEFAULT)))
-    if err:
-        delay = timedelta(minutes=RETRY_DELAY_MIN)
-    with _LOCK:
-        _STATE["last_scan"] = now.strftime("%Y-%m-%d %H:%M:%S")
-        _STATE["next_scan"] = (now + delay).strftime("%Y-%m-%d %H:%M:%S")
-        _STATE["last_error"] = err
-        _save_locked()
-    out["error"] = err
-    out["ok"] = not err
-    return out
-
-
-def fire_due():
-    """automation._tick 每拍调用：内部自节流，没到点/未启用时零开销返回。"""
-    _ensure_loaded()
-    with _LOCK:
-        if not (_STATE.get("config") or {}).get("poll_enabled"):
-            return None
-    return _poll(force=False)
-
-
-def scan_now():
-    """手动「立即扫描」：绕过轮询闸与节流。"""
-    return _poll(force=True)
-
-
-def _boot_reconcile():
-    """服务重启后补一次对账：上个进程死亡窗口里的终态变化别等下个 tick
-    （#27697 案：服务死窗 7 小时，失败转派被拖到重启后才发现）。"""
-    try:
-        _ensure_loaded()
-        with _LOCK:
-            cfg = _cfg()
-        _reconcile(cfg)
-    except Exception:
-        log.exception("zentao: 启动补对账失败")
-
-
-def start():
-    """服务启动接线：加载状态（含老配置迁移）。"""
-    global _BOOT_TIMER
-    n = load()
-    with _LOCK:
-        cfg = _STATE.get("config") or {}
-        if cfg.get("poll_enabled") and not _STATE.get("next_scan"):
-            _STATE["next_scan"] = time.strftime("%Y-%m-%d %H:%M:%S")
-            _save_locked()
-        pending = any(c.get("state") == "fixing" for c in _STATE["claims"].values())
-    if pending:
-        if _BOOT_TIMER is not None:
-            _BOOT_TIMER.cancel()
-        _BOOT_TIMER = threading.Timer(3.0, _boot_reconcile)
-        _BOOT_TIMER.daemon = True
-        _BOOT_TIMER.start()
-    return n
-
-
-# ---------------------------------------------------------------- 配置管理
-
-def _norm_profile(p, idx):
-    if not isinstance(p, dict):
-        raise ValueError("产品档案 #%d 必须是对象" % idx)
-    try:
-        pid = int(p.get("product") or 0)
-    except (TypeError, ValueError):
-        raise ValueError("产品档案 #%d 的产品 ID 必须是整数" % idx)
-    if pid <= 0:
-        raise ValueError("产品档案 #%d 缺产品 ID" % idx)
-    prof = dict(_PROFILE_DEFAULTS)
-    prof["product"] = pid
-    prof["assigned_to"] = str(p.get("assigned_to") or "").strip()
-    try:
-        prof["severity_cap"] = max(0, min(4, int(p.get("severity_cap") or 0)))
-    except (TypeError, ValueError):
-        prof["severity_cap"] = 0
-    our = [s for s in SIDES if s in (p.get("our_sides") or [])]
-    prof["our_sides"] = our
-    repos = {"backend": dict(_REPO_DEFAULTS), "frontend": dict(_REPO_DEFAULTS)}
-    raw_repos = p.get("repos") if isinstance(p.get("repos"), dict) else {}
-    for side in SIDES:
-        r = raw_repos.get(side) if isinstance(raw_repos.get(side), dict) else {}
-        wd = str(r.get("workdir") or "").strip()
-        if wd:
-            rp = Path(wd).expanduser()
-            if not rp.is_absolute():
-                raise ValueError("产品 %d 的%s工作目录必须是绝对路径" % (pid, SIDE_CN[side]))
-            wd = str(rp)
-        repos[side] = {"workdir": wd,
-                       "git_rev": str(r.get("git_rev") or "").strip(),
-                       "verify_command": str(r.get("verify_command") or "").strip()}
-    # 工作目录留空合法：扫描时回落「默认保存路径」（_launch_fix 同口径），不在此拦截
-    prof["repos"] = repos
-    hints = {}
-    raw_hints = p.get("repo_hints") if isinstance(p.get("repo_hints"), dict) else {}
-    for side in SIDES:
-        hints[side] = str(raw_hints.get(side) or "").strip()
-    prof["repo_hints"] = hints
-    owners = {}
-    raw_owners = p.get("owners") if isinstance(p.get("owners"), dict) else {}
-    for k in ("backend", "frontend", "not_ours"):
-        owners[k] = str(raw_owners.get(k) or "").strip()
-    prof["owners"] = owners
-    routes = []
-    for r in (p.get("module_routes") or []):
-        if not isinstance(r, dict):
-            continue
-        try:
-            mid = int(r.get("module") or 0)
-        except (TypeError, ValueError):
-            raise ValueError("产品 %d 的模块路由：模块 ID 必须是整数" % pid)
-        if mid <= 0:
-            continue
-        side = str(r.get("side") or "")
-        if side not in TRIAGE_SIDES:
-            raise ValueError("产品 %d 的模块路由 side 必须是 %s 之一"
-                             % (pid, "/".join(TRIAGE_SIDES)))
-        routes.append({"module": mid, "side": side,
-                       "account": str(r.get("account") or "").strip()})
-    prof["module_routes"] = routes
-    return prof
-
-
-def _norm_profiles(v):
-    if v is None:
-        return None
-    if not isinstance(v, list):
-        raise ValueError("product_profiles 必须是数组")
-    out, seen = [], set()
-    for i, p in enumerate(v):
-        prof = _norm_profile(p, i)
-        if prof["product"] in seen:
-            raise ValueError("产品 %d 配置重复" % prof["product"])
-        seen.add(prof["product"])
-        out.append(prof)
-    return out
-
-
-def save_config(patch):
-    """部分更新配置（password 缺省或空串=不改；product_profiles 整体替换）。
-    校验失败抛 ValueError。返回脱敏视图的 config。"""
-    _ensure_loaded()
-    patch = patch if isinstance(patch, dict) else {}
-    with _LOCK:
-        cfg = _cfg()
-        for k in _UPDATABLE:
-            if k not in patch:
-                continue
-            v = patch[k]
-            if k == "base_url":
-                v = str(v or "").strip().rstrip("/")
-                if v and not v.startswith(("http://", "https://")):
-                    raise ValueError("禅道地址必须以 http:// 或 https:// 开头")
-            elif k == "account":
-                v = str(v or "").strip()
-            elif k == "password":
-                v = str(v or "")
-                if not v:
-                    continue
-            elif k == "product_profiles":
-                v = _norm_profiles(v)
-                if v is None:
-                    continue
-            elif k == "interval_minutes":
-                try:
-                    v = max(INTERVAL_MIN, min(INTERVAL_MAX, int(v)))
-                except (TypeError, ValueError):
-                    raise ValueError("interval_minutes 必须是 %d-%d 的整数"
-                                     % (INTERVAL_MIN, INTERVAL_MAX))
-            elif k in ("auto_resolve", "auto_merge", "triage_ai", "poll_enabled"):
-                v = bool(v)
-            cfg[k] = v
-        _STATE["config"] = cfg
-        if cfg.get("poll_enabled"):
-            _STATE["next_scan"] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-        _save_locked()
-    return view()["config"]
-
-
-def bug_web_base(cfg=None):
-    """给前端拼 bug 详情页用的 web 根：探测出的有效地基（含自动补的 /zentao 子路径）
-    优先，没探测过回落用户填的原始地址。返回 stripped 根或空串。"""
-    c = dict(cfg) if cfg else _cfg()
-    hit = resolved_base_url(c.get("base_url"))
-    if hit:
-        return hit
-    b = re.sub(r"/api\.php/v1$", "", str(c.get("base_url") or "").strip().rstrip("/"))
-    return b if b.startswith(("http://", "https://")) else ""
-
-
-def bug_link_style(cfg=None):
-    """bug 详情链接形态：探测缓存优先，没探出来按伪静态（官方默认形态）。"""
-    c = dict(cfg) if cfg else _cfg()
-    b = str(c.get("base_url") or "").strip().rstrip("/")
-    return _WEBSTYLE.get(b) or "pathinfo"
-
-
-def view():
-    """前端视图：配置脱敏（password 只回是否已设）+ claims 列表（新在前）。"""
-    _ensure_loaded()
-    with _LOCK:
-        cfg = dict(_cfg())
-        has_pw = bool(cfg.get("password"))
-        cfg["password"] = ""
-        cfg["has_password"] = has_pw
-        bug_base = bug_web_base(cfg)
-        claims = sorted(_STATE["claims"].values(),
-                        key=lambda c: str(c.get("claimed_at") or ""), reverse=True)
-        return {"config": cfg,
-                "bug_base": bug_base,
-                "bug_style": bug_link_style(cfg),
-                "claims": [dict(c) for c in claims],
-                "last_scan": _STATE.get("last_scan") or "",
-                "next_scan": _STATE.get("next_scan") or "",
-                "last_error": _STATE.get("last_error") or ""}
-
-
-def _test_reset():
-    """测试钩子：清空内存状态（配合重绑 _FILE + TUTTI_DATA 临时目录用）。"""
-    with _LOCK:
-        _STATE["config"] = dict(_CFG_DEFAULTS)
-        _STATE["claims"] = {}
-        _STATE["last_scan"] = _STATE["next_scan"] = _STATE["last_error"] = ""
-        _reset_token()
-        _APIBASE.clear()
-        _WEBSTYLE.clear()
-        _MODE.clear()
-        _OLD.update(api="", sid="", at=0.0)
-        _OLD_FORM["form"] = ""
-        _REST_LAST_ERR["msg"] = ""
-        _IMG_CACHE.clear()
-        globals()["_LOADED"] = True
+# -*- coding: utf-8 -*-
+"""禅道 Bug 自动修复对接：多产品档案 + 排查定责路由 + 转派流转。
+
+数据落盘 <data>/zentao.json（tmp + os.replace 原子写；TUTTI_DATA 环境变量感知）。
+修复走与 /api/tasks 完全相同的链路（store.create_task → store.create_run →
+jobs.enqueue，code 引擎=实现→验证→评审→修复），不自造运行器。调度挂在
+automation._tick（同 publish/auto.fire_due 模式），内部按 interval_minutes 节流。
+
+禅道 REST API v1（开源版 15.x+；请求头 Token: <token>）：
+  POST {base}/api.php/v1/tokens               {account, password} → {token}
+  GET  {base}/api.php/v1/products/{pid}/bugs  分页 {bugs:[...], page, total, limit}
+  GET  {base}/api.php/v1/bugs/{id}            单查（回写前确认状态防谎报）
+  POST {base}/api.php/v1/bugs/{id}/resolve    {resolution, resolvedBuild, comment, assignedTo}
+  PUT  {base}/api.php/v1/bugs/{id}            {assignedTo, comment}（转派/失败说明）
+  GET  {base}/api.php/v1/products|users       产品/账号清单（前端下拉辅助）
+
+地址自适应：根路径 tokens 404 时自动试 {base}/zentao/api.php/v1（官方一键安装包
+默认子路径部署），探测成功缓存在 _APIBASE（key=用户填的原始地址）。
+
+产品档案（product_profiles）：每产品一份 {指派过滤, 严重度, 我方端 our_sides,
+后端/前端仓库(workdir/git_rev/verify_command), repo_hints, 负责人 owners,
+模块路由 module_routes}。老版扁平配置（products+单仓库）load 时自动迁移。
+
+排查（_triage）：模块路由按 bug.module 精确匹配优先 → AI 兜底（triage_ai 开时
+modelhub 单次调用，bookmeta 同款配方）→ unknown。判定 side ∈
+backend | frontend | both | not_ours | unknown。
+
+流转规则（转派目标只认排查结论，与 bug 当前 assignedTo 无关——测试提错人也
+照样改派）：
+  我方端问题    → 建修复任务；成功=合并+resolve(fixed)+报告评论+指回报告人
+  双端/我方一端 → 我方端修完（合并落库）后转派另一端负责人+评论，不 resolve
+  纯对方端问题  → 不建任务，直接转派该端负责人+排查结论评论
+  非我方        → 转派报告人（或 owners.not_ours）+评论；只转派不解决
+  unknown       → 不碰 bug，need_manual + 群通知（下轮扫描 bug 仍激活则重排查）
+  修复任务失败  → 评论尝试记录 + 转派该端负责人（模块路由 account > 端负责人）；
+                  CodeBee 自身崩溃（内部异常）不转派——那是工具故障不是修不动，
+                  只评论+群通知人工；负责人未配置同样只评论。
+                  只评论也绝不盲写 PUT：禅道 PUT /bugs/{id} 缺 assignedTo 会
+                  清空指派人，必须带回当前指派人（#27783 案）。
+
+落库纪律：修复任务落单即带基线（档案配了用档案，没配取工作目录当前 HEAD）
+走任务分支隔离，修完的代码自动提交在任务分支上，对账合并后才算落库；
+resolve 前再验一遍——没基线的任务改动只在工作区，有未提交的已跟踪改动
+不 resolve（转 done_manual 人工收口），auto_merge 关闭同理。
+对账（_reconcile）认任务最新一次运行：人工重试换 run 不影响回写判断。
+
+出网边界（SSRF 防护，_guard_url）：请求目标来自用户自配禅道地址——内网按设计
+放行；强制 http(s)、解析主机并阻断云元数据/链路本地地址、禁跟随重定向。
+"""
+from __future__ import annotations
+
+import base64
+import html as _html
+import hashlib
+import io
+import ipaddress
+import json
+import logging
+import os
+import re
+import socket
+import threading
+import time
+import urllib.error
+import urllib.parse
+import urllib.request
+from datetime import datetime, timedelta
+from pathlib import Path
+
+from . import jobs, operations, paths, settings, store, tlsctx
+
+log = logging.getLogger(__name__)
+
+_LOCK = threading.RLock()
+_SCAN_LOCK = threading.Lock()  # 手动/定时扫描单飞，避免重复拉取和重复认领
+_FILE = paths.DATA_DIR / "zentao.json"
+_STATE = {
+    "config": {},        # 持久配置（_CFG_DEFAULTS）
+    "claims": {},        # str(bug_id) → claim dict（v2：含 triage/tasks）
+    "last_scan": "",
+    "next_scan": "",
+    "last_error": "",
+}
+_LOADED = False
+_BOOT_TIMER = None     # 启动补对账的兜底 Timer（测试里要能拿到并取消）
+
+TOKEN_TTL = 23 * 3600
+HTTP_TIMEOUT = 15
+PAGE_LIMIT = 100
+MAX_BUGS = 500
+RESOLVE_MAX_ATTEMPTS = 3
+RETRY_DELAY_MIN = 30
+INTERVAL_MIN, INTERVAL_MAX = 5, 10080   # 扫描间隔（分钟）：5 分钟 ~ 7 天
+INTERVAL_DEFAULT = 5
+_LEGACY_INTERVAL_HOURS_DEFAULT = 2      # 老配置 interval_hours 缺省值（迁移用）
+
+SIDES = ("backend", "frontend")
+TRIAGE_SIDES = ("backend", "frontend", "both", "not_ours")   # unknown 单列
+SIDE_CN = {"backend": "后端", "frontend": "前端"}
+
+# 图片证据：#27754 案——「功能未对齐，具体如图」被纯文本排查脑补成前端样式问题。
+# 排查/修复前先尽量把图抓到手；抓不到就按守卫留人工，绝不盲判。
+_IMG_REF_RE = re.compile(r"如图|见图|截图|附图|下图|上图|图示|图片|图像|录屏|视频")
+TRIAGE_MAX_IMAGES = 4          # 排查提示词最多随附几张图（请求体与 token 预算）
+IMG_CACHE_TTL = 900            # bug 图片按 bug_id 短缓存（排查+建任务两处共用）
+_IMG_CACHE = {}                # str(bug_id) → (ts, [(mime, b64)])；空结果也缓存防重复下载
+
+_CFG_DEFAULTS = {
+    "base_url": "",
+    "account": "",
+    "password": "",
+    "product_profiles": [],    # 产品档案列表（_norm_profile 形状）
+    "auto_resolve": True,
+    "auto_merge": True,
+    "triage_ai": True,         # 模块路由未命中时用 AI 兜底排查
+    "poll_enabled": False,
+    "interval_minutes": INTERVAL_DEFAULT,
+}
+
+_REPO_DEFAULTS = {"workdir": "", "git_rev": "", "verify_command": ""}
+
+_PROFILE_DEFAULTS = {
+    "product": 0,              # 产品 ID（必填唯一）
+    "assigned_to": "",         # 只认领指派给该账号的 bug；空=不按指派过滤
+    "severity_cap": 0,         # 严重度上限（1 最严重）；0=不限
+    "our_sides": ["backend"],  # 我方端（CodeBee 自动修）；空=纯排查转派不修
+    "repos": {"backend": dict(_REPO_DEFAULTS), "frontend": dict(_REPO_DEFAULTS)},
+    "repo_hints": {"backend": "", "frontend": ""},   # AI 排查时的一句仓库描述
+    "owners": {"backend": "", "frontend": "", "not_ours": ""},
+    "module_routes": [],       # [{module:int, side:backend|frontend|both|not_ours, account:""}]
+}
+
+_UPDATABLE = ("base_url", "account", "password", "product_profiles",
+              "auto_resolve", "auto_merge", "triage_ai",
+              "poll_enabled", "interval_minutes")
+
+
+# ---------------------------------------------------------------- 出网边界（SSRF）
+
+_META_HOSTS = {"metadata.google.internal", "metadata.goog"}
+_NO_REDIRECT = None
+
+
+def _no_redirect_opener():
+    """HTTP 重定向一概不跟：目标必须就是用户配置的那台禅道。"""
+    global _NO_REDIRECT
+    if _NO_REDIRECT is None:
+        class _Stop(urllib.request.HTTPRedirectHandler):
+            def redirect_request(self, req, fp, code, msg, headers, newurl):
+                return None
+        _NO_REDIRECT = urllib.request.build_opener(
+            _Stop(), urllib.request.HTTPSHandler(context=tlsctx.context()))
+    return _NO_REDIRECT
+
+
+def _guard_url(url):
+    """出网前的边界校验：协议白名单 + 主机解析阻断云元数据/链路本地地址。
+
+    私网与环回按设计放行（禅道自建在内网是主场景，配置者即本机用户）。
+    校验失败抛 ZenError。返回原 url。
+    """
+    from urllib.parse import urlparse
+    u = urlparse(url)
+    if u.scheme not in ("http", "https"):
+        raise ZenError("禅道地址只允许 http/https 协议")
+    host = (u.hostname or "").lower()
+    if not host:
+        raise ZenError("禅道地址缺少主机名")
+    if host in _META_HOSTS:
+        raise ZenError("不允许访问云元数据地址")
+    try:
+        infos = socket.getaddrinfo(host, None)
+    except OSError:
+        raise ZenError("禅道主机解析失败：%s" % host)
+    for info in infos:
+        ip = ipaddress.ip_address(info[4][0])
+        if ip.is_link_local or ip.is_reserved and not ip.is_private:
+            raise ZenError("不允许访问链路本地/保留地址：%s" % ip)
+        if ip.is_multicast or ip.is_unspecified:
+            raise ZenError("不允许访问多播/未指定地址：%s" % ip)
+    return url
+
+
+# ---------------------------------------------------------------- 持久化
+
+def _save_locked():
+    _FILE.parent.mkdir(parents=True, exist_ok=True)
+    tmp = _FILE.with_suffix(".tmp")
+    tmp.write_text(json.dumps({"version": 2, **_STATE},
+                              ensure_ascii=False, indent=2), encoding="utf-8")
+    os.replace(str(tmp), str(_FILE))
+
+
+def _normalize_claim(d):
+    """claim 归一：v1 单任务形状（task_id/run_id）→ v2 tasks 数组。"""
+    c = dict(d) if isinstance(d, dict) else {}
+    if isinstance(c.get("tasks"), list) and c["tasks"]:
+        tasks = [dict(t) for t in c["tasks"] if isinstance(t, dict)]
+    else:
+        tasks = [{"side": "backend", "task_id": c.get("task_id") or "",
+                  "run_id": c.get("run_id") or "", "state": "fixing"}] \
+            if (c.get("task_id") or c.get("run_id")) else []
+    c["tasks"] = tasks
+    c.setdefault("triage", {"side": "", "reason": "", "by": "", "account": ""})
+    if not isinstance(c.get("triage"), dict):
+        c["triage"] = {"side": "", "reason": "", "by": "", "account": ""}
+    c.setdefault("note", "")
+    c.setdefault("attempts", 0)
+    return c
+
+
+def _migrate_legacy(cfg):
+    """老扁平配置（products+单仓库）→ product_profiles。就地改写返回。"""
+    if cfg.get("product_profiles"):
+        return cfg
+    products = cfg.get("products") or []
+    if not isinstance(products, list) or not products:
+        return cfg
+    try:
+        pids = [int(p) for p in products if str(p).strip()]
+    except (TypeError, ValueError):
+        return cfg
+    backend = dict(_REPO_DEFAULTS)
+    for k in _REPO_DEFAULTS:
+        backend[k] = str(cfg.get(k) or "")
+    our = ["backend"]
+    if not backend["workdir"]:
+        our = []          # 老配置连工作目录都没配：纯路由
+    for pid in pids:
+        cfg.setdefault("product_profiles", []).append({
+            "product": pid,
+            "assigned_to": str(cfg.get("assigned_to") or ""),
+            "severity_cap": int(cfg.get("severity_cap") or 0),
+            "our_sides": list(our),
+            "repos": {"backend": backend, "frontend": dict(_REPO_DEFAULTS)},
+            "repo_hints": {"backend": "", "frontend": ""},
+            "owners": {"backend": "", "frontend": "", "not_ours": ""},
+            "module_routes": [],
+        })
+    return cfg
+
+
+def load(force=False):
+    """从磁盘加载状态（幂等）。返回 claims 数。测试可重绑 _FILE 后 force=True。"""
+    global _LOADED
+    with _LOCK:
+        if _LOADED and not force:
+            return len(_STATE["claims"])
+        try:
+            data = json.loads(_FILE.read_text(encoding="utf-8"))
+        except Exception:
+            data = {}
+        cfg = data.get("config") if isinstance(data, dict) else None
+        # 老扁平配置先迁移成产品档案（老键 products/workdir/... 不在新 defaults 里，
+        # 必须在按 _CFG_DEFAULTS 过滤之前完成迁移）
+        if isinstance(cfg, dict) and not cfg.get("product_profiles"):
+            cfg = _migrate_legacy(dict(cfg))
+        # 老配置用 interval_hours（小时），新配置改用 interval_minutes（分钟）：
+        # 用户没动过间隔（还是老默认 2 小时）就跟随新默认 5 分钟；真改过的按小时换算保留。
+        if isinstance(cfg, dict) and "interval_minutes" not in cfg \
+                and "interval_hours" in cfg:
+            try:
+                hours = int(cfg.get("interval_hours") or _LEGACY_INTERVAL_HOURS_DEFAULT)
+            except (TypeError, ValueError):
+                hours = _LEGACY_INTERVAL_HOURS_DEFAULT
+            cfg = dict(cfg)
+            mins = INTERVAL_DEFAULT if hours == _LEGACY_INTERVAL_HOURS_DEFAULT else hours * 60
+            cfg["interval_minutes"] = max(INTERVAL_MIN, min(INTERVAL_MAX, mins))
+        merged = dict(_CFG_DEFAULTS)
+        if isinstance(cfg, dict):
+            merged.update({k: v for k, v in cfg.items() if k in _CFG_DEFAULTS})
+        claims = data.get("claims") if isinstance(data, dict) else None
+        _STATE["config"] = merged
+        _STATE["claims"] = {str(k): _normalize_claim(v)
+                            for k, v in (claims or {}).items()} \
+            if isinstance(claims, dict) else {}
+        _STATE["last_scan"] = str(data.get("last_scan") or "") if isinstance(data, dict) else ""
+        _STATE["next_scan"] = str(data.get("next_scan") or "") if isinstance(data, dict) else ""
+        _STATE["last_error"] = str(data.get("last_error") or "") if isinstance(data, dict) else ""
+        _LOADED = True
+        return len(_STATE["claims"])
+
+
+def _ensure_loaded():
+    if not _LOADED:
+        load()
+
+
+def _cfg():
+    cfg = dict(_CFG_DEFAULTS)
+    cfg.update({k: v for k, v in (_STATE.get("config") or {}).items()
+                if k in _CFG_DEFAULTS})
+    return cfg
+
+
+def _profiles(cfg=None):
+    out = []
+    for p in (cfg or _cfg()).get("product_profiles") or []:
+        prof = dict(_PROFILE_DEFAULTS)
+        prof.update({k: v for k, v in (p or {}).items() if k in _PROFILE_DEFAULTS})
+        repos = dict(_PROFILE_DEFAULTS["repos"])
+        for side in SIDES:
+            r = dict(_REPO_DEFAULTS)
+            r.update({k: v for k, v in ((prof.get("repos") or {}).get(side) or {}).items()
+                      if k in _REPO_DEFAULTS})
+            repos[side] = r
+        prof["repos"] = repos
+        hints = dict(_PROFILE_DEFAULTS["repo_hints"])
+        for side in SIDES:
+            hints[side] = str((prof.get("repo_hints") or {}).get(side) or "")
+        prof["repo_hints"] = hints
+        owners = dict(_PROFILE_DEFAULTS["owners"])
+        for k in owners:
+            owners[k] = str((prof.get("owners") or {}).get(k) or "").strip()
+        prof["owners"] = owners
+        prof["our_sides"] = [s for s in SIDES if s in (prof.get("our_sides") or [])]
+        prof["module_routes"] = [r for r in (prof.get("module_routes") or [])
+                                 if isinstance(r, dict)]
+        out.append(prof)
+    return out
+
+
+def _profile_for(cfg, product_id):
+    try:
+        pid = int(product_id or 0)
+    except (TypeError, ValueError):
+        return None
+    for p in _profiles(cfg):
+        if p["product"] == pid:
+            return p
+    return None
+
+
+# ---------------------------------------------------------------- 禅道客户端
+
+class ZenError(Exception):
+    """禅道接口错误（message 人话，可直接展示）。"""
+
+
+_TOKEN = {"v": "", "at": 0.0}
+# 原始地址 → 探测成功的 api 基址（一键安装包常部署在 /zentao 子路径下，自动补探）
+_APIBASE = {}
+# 通道自适应：原始地址 → "rest"（≥15 REST v1）/ "old"（老版 module-method JSON 接口）
+_MODE = {}
+# web 端路由形态：原始地址 → "pathinfo"（伪静态 bug-view-N.html，官方默认）
+# / "get"（index.php?m=bug&f=view&bugID=N）。bug 详情链接用（2026-09-23 真机实证
+# 伪静态部署对 GET 式链接不路由）。
+_WEBSTYLE = {}
+_OLD = {"api": "", "sid": "", "at": 0.0}
+_OLD_FORM = {"form": ""}      # 命中的密码形态 "plain" | "md5chain"，缓存避免重复试
+_REST_LAST_ERR = {"msg": ""}
+_OLD_TTL = 20 * 60
+
+
+def _reset_token():
+    _TOKEN["v"] = ""
+    _TOKEN["at"] = 0.0
+
+
+def resolved_base_url(base_url):
+    """探测成功后的有效地基（含自动补出的子路径）；没探测过返回空串。"""
+    b = str(base_url or "").strip().rstrip("/")
+    hit = _APIBASE.get(b) or (_OLD.get("api") if _MODE.get(b) == "old" else "")
+    if not hit:
+        return ""
+    return re.sub(r"/api\.php/v1$", "", hit)
+
+
+def probe_web_style(base_url):
+    """探测禅道 web 端路由形态（bug 详情链接用）：GET web 根（裸根 + 常见 /zentao
+    子路径都试），看登录跳转里是伪静态路由（user-login-xxx.html）还是 GET 式
+    （index.php?m=user）。尽力而为：任何网络/守卫异常回空串，绝不影响登录与
+    扫描主流程；探出即缓存。"""
+    try:
+        b = _raw_base(base_url)
+    except ZenError:
+        return ""
+    if _WEBSTYLE.get(b):
+        return _WEBSTYLE[b]
+    style = ""
+    for root in (b + "/", b + "/zentao/"):
+        body = ""
+        try:
+            req = urllib.request.Request(_guard_url(root), headers={"Accept": "text/html"})
+            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
+                body = (r.read(65536) or b"").decode("utf-8", "replace")
+        except Exception:
+            body = ""
+        if "user-login-" in body or re.search(r"[a-z]+-view-\d+[^\"]*\.html", body):
+            style = "pathinfo"
+        elif "index.php?m=" in body:
+            style = "get"
+        if style:
+            break
+    if style:
+        _WEBSTYLE[b] = style
+    return style
+
+
+def _raw_base(base_url):
+    b = str(base_url or "").strip().rstrip("/")
+    if not b:
+        raise ZenError("禅道地址未配置")
+    if not b.startswith(("http://", "https://")):
+        raise ZenError("禅道地址必须以 http:// 或 https:// 开头")
+    return b
+
+
+def _api_base(base_url):
+    b = _raw_base(base_url)
+    hit = _APIBASE.get(b)
+    if hit:
+        return hit
+    return b + "/api.php/v1"
+
+
+def _base_candidates(b):
+    """api 基址候选：用户填的根路径 → 官方一键安装包常见的 /zentao 子路径。"""
+    out, seen = [], set()
+    for c in (b + "/api.php/v1", b + "/zentao/api.php/v1"):
+        if c not in seen:
+            seen.add(c)
+            out.append(c)
+    return out
+
+
+def _rest_login(base_url, account, password):
+    """REST v1 通道取 token。成功返回 token（缓存基址）；
+    通道不可用（404 / 只认应用 code 的 200 信封）返回 None；
+    账密被 REST 明确拒绝（401/403）或网络/守卫失败抛 ZenError。"""
+    b = _raw_base(base_url)
+    for api in _base_candidates(b):
+        url = _guard_url(api + "/tokens")
+        body = json.dumps({"account": str(account or ""), "password": str(password or "")},
+                          ensure_ascii=False).encode("utf-8")
+        req = urllib.request.Request(url, data=body, method="POST", headers={
+            "Content-Type": "application/json", "Accept": "application/json"})
+        try:
+            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
+                data = json.loads((r.read() or b"{}").decode("utf-8", "replace"))
+        except urllib.error.HTTPError as e:
+            try:
+                detail = json.loads((e.read() or b"").decode("utf-8", "replace"))
+                msg = (detail.get("error") or "") if isinstance(detail, dict) else ""
+            except Exception:
+                msg = ""
+            if e.code in (401, 403):
+                raise ZenError("禅道账号或密码不对（%s）%s" % (e.code, msg))
+            if e.code == 404:
+                _REST_LAST_ERR["msg"] = "REST 接口不存在（404）"
+                continue        # 换下一个候选基址
+            raise ZenError("禅道返回 %s：%s" % (e.code, msg or "获取令牌失败"))
+        except ZenError:
+            raise
+        except Exception as e:
+            raise ZenError("连不上禅道（%s）——请检查地址与网络" % e)
+        tok = ""
+        if isinstance(data, dict):
+            tok = str(data.get("token") or "")
+            if not tok and isinstance(data.get("data"), dict):
+                tok = str(data["data"].get("token") or "")
+        if tok:
+            _APIBASE[b] = api
+            _TOKEN["v"] = tok
+            _TOKEN["at"] = time.time()
+            probe_web_style(b)      # bug 链接形态探测，失败不影响登录
+            return tok
+        # HTTP 200 但没有 token：REST 在但不认账密（如只认应用 code 的部署）
+        _REST_LAST_ERR["msg"] = "REST 接口不认账密（%s）" % (
+            str(data.get("errmsg") or data.get("error") or "响应里没有 token"))
+    return None
+
+
+def _token(cfg, force=False):
+    if force or not _TOKEN["v"] or time.time() - _TOKEN["at"] > TOKEN_TTL:
+        tok = _rest_login(cfg.get("base_url"), cfg.get("account"), cfg.get("password"))
+        if not tok:
+            raise ZenError(_REST_LAST_ERR.get("msg") or "REST 通道不可用")
+        return tok
+    return _TOKEN["v"]
+
+
+def _api(method, path, cfg=None, body=None):
+    """调禅道 API。返回 dict（非 dict 响应返回 {}）。401/403 自动重取 token 重试一次。"""
+    c = cfg or _cfg()
+    for attempt in (1, 2):
+        headers = {"Accept": "application/json", "Token": _token(c, force=(attempt == 2))}
+        url = _guard_url(_api_base(c.get("base_url")) + path)
+        data = None
+        if body is not None:
+            data = json.dumps(body, ensure_ascii=False).encode("utf-8")
+            headers["Content-Type"] = "application/json"
+        req = urllib.request.Request(url, data=data, method=method, headers=headers)
+        try:
+            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
+                out = json.loads((r.read() or b"{}").decode("utf-8", "replace"))
+            return out if isinstance(out, dict) else {"_list": out}
+        except urllib.error.HTTPError as e:
+            raw = ""
+            try:
+                raw = (e.read() or b"").decode("utf-8", "replace")
+            except Exception:
+                pass
+            detail = ""
+            try:
+                d = json.loads(raw)
+                if isinstance(d, dict):
+                    detail = str(d.get("error") or d.get("message") or "")
+            except Exception:
+                pass
+            if e.code in (401, 403) and attempt == 1:
+                _reset_token()
+                continue
+            raise ZenError("禅道接口 %s（%s）%s" % (path, e.code, detail or "调用失败"))
+        except ZenError:
+            raise
+        except Exception as e:
+            raise ZenError("连不上禅道（%s）" % e)
+    raise ZenError("禅道认证失败（token 两次获取后仍被拒绝，检查账号权限）")
+
+
+# ---------------------------------------------------------------- 老版 JSON 接口适配
+# 禅道 15 以前的经典入口：/zentao/api-getsessionid.json 取会话 → user-login.json
+# 账密登录（Cookie zentaosid）→ {module}-{method}-{参数}.json 调业务。响应是双层
+# 信封 {"status","data":"<json 字符串>"}。有的部署 REST 只认应用 code，账密只能
+# 走这条通道，故在 _call 里自动探测分路。REST 路径在这里翻译成老接口形态。
+
+def _old_parse(raw):
+    """老信封解析。非 JSON（多半是登录重定向 HTML）返回 None = 会话死/不可用。"""
+    try:
+        outer = json.loads(raw.decode("utf-8", "replace"))
+    except Exception:
+        return None
+    if isinstance(outer, dict) and isinstance(outer.get("data"), str):
+        try:
+            outer["data"] = json.loads(outer["data"])
+        except Exception:
+            pass
+    return outer if isinstance(outer, dict) else None
+
+
+def _old_fail(outer):
+    """status=failed / result=fail → 人话错误文本；成功返回空串。"""
+    if not isinstance(outer, dict):
+        return ""
+    if outer.get("status") == "failed" or outer.get("result") == "fail":
+        return str(outer.get("reason") or outer.get("message")
+                   or outer.get("error") or "调用失败")
+    return ""
+
+
+def _old_get(api, path):
+    url = _guard_url(api + path)
+    req = urllib.request.Request(url, headers={"Accept": "application/json"})
+    with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
+        return r.read() or b""
+
+
+def _old_post(api, path, form):
+    url = _guard_url(api + path)
+    data = urllib.parse.urlencode(form or {}).encode("utf-8")
+    req = urllib.request.Request(url, data=data, method="POST", headers={
+        "Content-Type": "application/x-www-form-urlencoded",
+        "Accept": "application/json"})
+    with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
+        return r.read() or b""
+
+
+def _old_session(api):
+    """取新会话。返回 (sid, rand)；通道不可用返回 None。"""
+    try:
+        raw = _old_get(api, "/api-getsessionid.json")
+    except urllib.error.HTTPError:
+        return None
+    except ZenError:
+        raise
+    except Exception as e:
+        raise ZenError("连不上禅道（%s）——请检查地址与网络" % e)
+    outer = _old_parse(raw)
+    if not outer or not isinstance(outer.get("data"), dict):
+        return None
+    d = outer["data"]
+    return str(d.get("sessionID") or ""), str(d.get("rand") or "")
+
+
+def _old_probe_authed(api, sid):
+    """登录成功的兜底判据：带会话调需登录端点拿得到 JSON 信封（未登录会被
+    重定向到登录页 HTML，解析出来是 None）。"""
+    try:
+        raw = _old_get(api, "/my-index.json?zentaosid=" + sid)
+    except Exception:
+        return False
+    return _old_parse(raw) is not None
+
+
+def _old_identify(api, sid, rand, account, password, form):
+    """老接口账密登录。返回 (ok, 人话错误)。form=plain|md5chain。"""
+    pw = str(password or "")
+    if form == "md5chain":
+        pw = hashlib.md5((hashlib.md5(pw.encode("utf-8")).hexdigest()
+                          + str(rand)).encode("utf-8")).hexdigest()
+    try:
+        raw = _old_post(api, "/user-login.json?zentaosid=" + sid,
+                        {"account": str(account or ""), "password": pw})
+    except Exception as e:
+        return False, "连不上禅道（%s）" % e
+    outer = _old_parse(raw)
+    if outer is None:
+        return False, "登录响应不可识别"
+    err = _old_fail(outer)
+    if err:
+        return False, err
+    # 成功响应形状各版本不一：user 可能在 data 里（老）或顶层（实测某老版部署）
+    d = outer.get("data")
+    if (isinstance(d, dict) and isinstance(d.get("user"), dict)) \
+            or isinstance(outer.get("user"), dict) or _old_probe_authed(api, sid):
+        return True, ""
+    return False, "登录未被接受（账号或密码不对，或账号被锁）"
+
+
+def _old_login(cfg, rest_err=""):
+    """老接口通道登录（带 /zentao 子路径候选）。成功置 _OLD 缓存与密码形态。"""
+    b = _raw_base(cfg.get("base_url"))
+    account, password = cfg.get("account"), cfg.get("password")
+    cands, seen = [], set()
+    for c in (b + "/zentao", b):
+        if c not in seen:
+            seen.add(c)
+            cands.append(c)
+    last = ""
+    for api in cands:
+        sess = _old_session(api)
+        if not sess:
+            continue
+        sid, rand = sess
+        forms = [_OLD_FORM["form"] or "plain", "md5chain", "plain"]
+        tried = set()
+        for form in [f for f in forms if not (f in tried or tried.add(f))]:
+            ok, err = _old_identify(api, sid, rand, account, password, form)
+            if ok:
+                _OLD.update(api=api, sid=sid, at=time.time())
+                _OLD_FORM["form"] = form
+                _MODE[b] = "old"
+                probe_web_style(b)  # bug 链接形态探测，失败不影响登录
+                return
+            last = err
+        break       # 同一台服务，账密结果与子路径无关，别再烧尝试次数
+    if not last:
+        raise ZenError("禅道老版接口不可用（会话接口无响应——地址若是子目录部署要带上子目录）"
+                       + ("；%s" % rest_err if rest_err else ""))
+    if "账号" in last or "密码" in last or "锁定" in last:
+        raise ZenError(last)
+    raise ZenError("禅道老接口登录失败：%s" % last)
+
+
+def _old_call(method, path, cfg, body=None):
+    """老接口执行：翻译 REST 路径 → module-method.json，返回 REST 同形状的 dict。"""
+    b = _raw_base(cfg.get("base_url"))
+    if _MODE.get(b) != "old" or not _OLD["sid"] or time.time() - _OLD["at"] > _OLD_TTL:
+        _old_login(cfg)
+    parts = urllib.parse.urlsplit(path)
+    q = dict(urllib.parse.parse_qsl(parts.query))
+    p = parts.path
+    for attempt in (1, 2):
+        try:
+            return _old_route(_OLD["api"], method, p, q, body)
+        except _OldSessionDead:
+            if attempt == 2:
+                break
+            _OLD["sid"] = ""
+            _old_login(cfg)
+        except urllib.error.HTTPError as e:
+            raise ZenError("禅道老接口 %s（%s）调用失败" % (p, e.code))
+    raise ZenError("禅道会话两次登录后仍失效（检查账号权限）")
+
+
+class _OldSessionDead(Exception):
+    pass
+
+
+def _pairs_to_list(pairs, id_key, name_key):
+    """{id/name: 值} 形态的 pairs → [{id,name}] 列表（值可能为串或对象）。"""
+    out = []
+    for k, v in (pairs or {}).items():
+        if isinstance(v, dict):
+            item = dict(v)
+            item.setdefault(id_key, k)
+        else:
+            item = {id_key: k, name_key: v}
+        out.append(item)
+    return out
+
+
+def _module_items(d):
+    """老版模块树形状兼容（2026-09-21 真机反馈「响应形状不认识」）：
+    sons/modules 数组、{id: {name,...}} 字典、children/sons 嵌套树，统一
+    递归展开成 [{id,name}] 平铺清单（id=0 是树根容器，跳过）；全数字键的
+    {id: 名字} 扁平映射也认（部分版本的模块摘要形状）。"""
+    out = []
+
+    def walk(node):
+        if isinstance(node, list):
+            for x in node:
+                walk(x)
+        elif isinstance(node, dict):
+            if node.get("id") not in (None, "", 0):
+                out.append(node)
+                for k in ("children", "sons"):
+                    if isinstance(node.get(k), (list, dict)):
+                        walk(node[k])
+            else:
+                if node and all(str(k).isdigit() for k in node.keys()):
+                    flat = [{"id": k, "name": str(v)}
+                            for k, v in node.items() if str(v or "").strip()]
+                    if flat:
+                        out.extend(flat)
+                        return
+                for v in node.values():
+                    walk(v)
+
+    if isinstance(d, dict):
+        # tree 优先：真机 tree=带 children 的完整嵌套树，sons 只有顶层扁平表
+        # （走 sons 会漏子模块）
+        walk(d.get("tree") or d.get("sons") or d.get("modules") or d)
+    elif isinstance(d, list):
+        walk(d)
+    return out
+
+
+def _old_route(api, method, p, q, body):
+    """路径翻译 + 请求 + 归一。会话死抛 _OldSessionDead。"""
+
+    def go(raw, translator=None):
+        outer = _old_parse(raw)
+        if outer is None:
+            txt = raw.decode("utf-8", "replace")
+            if "user-login" in txt:
+                raise _OldSessionDead()     # 登录重定向 = 会话死
+            # 其余 HTML 是 js::locate/alert 回包（老禅道 POST 动作成功就回这种）
+            return {"_js": True}
+        err = _old_fail(outer)
+        if err:
+            raise ZenError("禅道老接口%s：%s" % (translator or "", err))
+        return outer.get("data")
+
+    def get(path):
+        try:
+            raw = _old_get(api, path + ("&" if "?" in path else "?")
+                           + "zentaosid=" + _OLD["sid"])
+        except urllib.error.HTTPError as e:
+            raw = e.read() or b""
+            if _old_parse(raw) is None:
+                raise _OldSessionDead()     # 非 JSON：多半被重定向到登录页
+        return raw
+
+    def post(path, form):
+        try:
+            raw = _old_post(api, path + ("&" if "?" in path else "?")
+                            + "zentaosid=" + _OLD["sid"], form)
+        except urllib.error.HTTPError as e:
+            raw = e.read() or b""
+            if e.code in (401, 403) or _old_parse(raw) is None:
+                raise _OldSessionDead()
+        return raw
+
+    m = re.match(r"^/products/(\d+)/bugs$", p)
+    if m and method == "GET":
+        pid = m.group(1)
+        page = max(1, int(q.get("page") or 1))
+        out, total, seen_ids = [], 0, set()
+        # 翻页用路径参数形态（实测部分老版不认 query 翻页参数）：带 branch 段；
+        # 若 recPerPage 没生效（老版本无 branch 段会错位解析）回落 query 形态。
+        branch_form = True
+        for pg in range(page, page + 6):     # 去重兜底：翻页参数不被认时不会死循环
+            if branch_form:
+                pathq = "/bug-browse-%s-0-unclosed-0-id_desc-0-%d-%d.json" % (pid, PAGE_LIMIT, pg)
+            else:
+                pathq = "/bug-browse-%s.json?browseType=unclosed&orderBy=id_desc" \
+                        "&recPerPage=%d&pagerID=%d" % (pid, PAGE_LIMIT, pg)
+            d = go(get(pathq), "（拉 bug 列表）")
+            if branch_form and isinstance(d, dict):
+                try:
+                    got = int((d.get("pager") or {}).get("recPerPage") or 0)
+                except (TypeError, ValueError):
+                    got = 0
+                if got != PAGE_LIMIT:
+                    branch_form = False
+                    continue                 # 形态没对上：换 query 形态重拉本页
+            if not isinstance(d, dict):
+                break
+            bugs = d.get("bugs") if isinstance(d.get("bugs"), list) else []
+            fresh = [x for x in bugs if isinstance(x, dict)
+                     and str(x.get("id")) not in seen_ids]
+            for x in fresh:
+                seen_ids.add(str(x.get("id")))
+            out.extend(fresh)
+            try:
+                total = int((d.get("pager") or {}).get("recTotal") or 0)
+            except (TypeError, ValueError, AttributeError):
+                total = 0
+            if not fresh or (total and len(out) >= total) or len(out) >= MAX_BUGS:
+                break
+        return {"bugs": out[:MAX_BUGS], "total": total or len(out)}
+
+    m = re.match(r"^/products/(\d+)/modules$", p)
+    if m and method == "GET":
+        # 2026-09-22 真机实探：该版模块行 type=story，viewType=module 查出空树
+        # （status 仍 success）——「响应形状不认识」的真根因。改按 bug 视图拉
+        # （bug 表单模块下拉用的就是这棵树），story/case 兜底。
+        pid = m.group(1)
+        last, saw_js = None, False
+        for vt in ("bug", "story", "case"):
+            d = go(get("/tree-browse-%s-%s.json" % (pid, vt)), "（拉模块清单）")
+            last = d
+            if isinstance(d, dict) and d.get("_js"):
+                saw_js = True            # 接口回了页面脚本：换视图再试
+                continue
+            items = [x for x in _module_items(d) if isinstance(x, dict) and x.get("id")]
+            if items:
+                return {"_list": items}
+        if saw_js:
+            raise ZenError("禅道老接口（拉模块清单）：回了页面脚本而不是数据"
+                           "（模块树接口不在或被重定向）——请在禅道产品视图 URL 里查模块 ID 手工填写")
+        keys = ",".join(sorted(str(k) for k in last.keys())) \
+            if isinstance(last, dict) else type(last).__name__
+        raise ZenError("该产品模块树为空（bug/story/case 三个视图都没拉到模块，"
+                       "响应顶层键：%s）——请到禅道确认产品下建过模块，或手工填模块 ID" % keys)
+
+    if p == "/products" and method == "GET":
+        d = go(get("/api-getmodel-product-getpairs.json"), "（拉产品清单）")
+        prods = d.get("products") if isinstance(d, dict) else None
+        if isinstance(prods, list):     # 部分老版回数组形状 [{id,name,...}]
+            return {"_list": [x for x in prods if isinstance(x, dict) and x.get("id")]}
+        pairs = prods if isinstance(prods, dict) else d
+        return {"_list": _pairs_to_list(pairs, "id", "name")}
+
+    if p == "/users" and method == "GET":
+        d = go(get("/api-getmodel-user-getpairs.json"), "（拉账号清单）")
+        return {"_list": _pairs_to_list(d, "account", "realname")}
+
+    m = re.match(r"^/bugs/(\d+)$", p)
+    if m and method == "GET":
+        d = go(get("/bug-view-%s.json" % m.group(1)), "（查 bug）")
+        bug = d.get("bug") if isinstance(d, dict) and isinstance(d.get("bug"), dict) else d
+        # 老接口的 files 附件在 data 顶层，不在 bug 里：合并进 bug dict（REST 形状兼容）
+        if isinstance(bug, dict) and isinstance(d, dict) \
+                and isinstance(d.get("files"), (dict, list)):
+            bug = dict(bug)
+            bug["files"] = d["files"]
+        return bug
+
+    m = re.match(r"^/bugs/(\d+)/resolve$", p)
+    if m and method == "POST":
+        form = {k: str(v or "") for k, v in (body or {}).items()}
+        go(post("/bug-resolve-%s.json" % m.group(1), form), "（resolve）")
+        return {"ok": True}
+
+    m = re.match(r"^/bugs/(\d+)$", p)
+    if m and method == "PUT":
+        form = {k: str(v or "") for k, v in (body or {}).items() if k in ("assignedTo", "comment")}
+        go(post("/bug-assignTo-%s.json" % m.group(1), form), "（转派）")
+        return {"ok": True}
+
+    raise ZenError("禅道老接口不认识该调用：%s %s（请升级禅道到 ≥15 用 REST 接口）"
+                   % (method, p))
+
+
+def _call(method, path, cfg=None, body=None):
+    """统一入口：REST v1 优先；通道探明后按模式分发（账密不被 REST 接受时自动
+    切老版 JSON 接口）。所有业务调用都走这里，拿到的都是 REST 形状。"""
+    c = cfg or _cfg()
+    b = _raw_base(c.get("base_url"))
+    mode = _MODE.get(b)
+    if not mode:
+        tok = _rest_login(b, c.get("account"), c.get("password"))
+        if tok:
+            _MODE[b] = mode = "rest"
+        else:
+            _old_login(c, _REST_LAST_ERR.get("msg", ""))
+            mode = "old"
+    if mode == "rest":
+        return _api(method, path, cfg=c, body=body)
+    return _old_call(method, path, c, body=body)
+
+
+def _acct(v):
+    """assignedTo/openedBy 兼容：新版是 {account,...} 用户对象，老版可能是
+    [账号, 姓名] 数组或账号串。"""
+    if isinstance(v, dict):
+        return str(v.get("account") or "").strip()
+    if isinstance(v, (list, tuple)) and v:
+        return str(v[0] or "").strip()
+    return str(v or "").strip()
+
+
+def _strip_html(raw):
+    """steps 字段是富文本 HTML：剥标签转纯文本。"""
+    txt = str(raw or "")
+    txt = re.sub(r"(?is)<(script|style).*?>.*?</\1>", " ", txt)
+    txt = re.sub(r"(?i)<br\s*/?>|</p>|</li>|</div>|</tr>", "\n", txt)
+    txt = re.sub(r"<[^>]+>", " ", txt)
+    txt = _html.unescape(txt)
+    txt = re.sub(r"[ \t\r]+", " ", txt)
+    txt = re.sub(r"\n\s*\n+", "\n", txt)
+    return txt.strip()
+
+
+def _severity(bug):
+    try:
+        return int(bug.get("severity") or 0)
+    except (TypeError, ValueError):
+        return 0
+
+
+# ---------------------------------------------------------------- bug 图片证据
+
+_IMG_MAGIC = ((b"\x89PNG", "image/png"), (b"\xff\xd8\xff", "image/jpeg"),
+              (b"GIF8", "image/gif"))
+
+
+def _img_mime(raw):
+    """按魔数判图片类型；非图返回空串（下载到错误页 HTML 时靠它拦住）。"""
+    raw = bytes(raw or b"")
+    for pre, mime in _IMG_MAGIC:
+        if raw.startswith(pre):
+            return mime
+    if raw[:4] == b"RIFF" and raw[8:12] == b"WEBP":
+        return "image/webp"
+    return ""
+
+
+def _img_ext_mime(ext):
+    return {"png": "image/png", "jpg": "image/jpeg", "jpeg": "image/jpeg",
+            "gif": "image/gif", "webp": "image/webp", "bmp": "image/bmp"}.get(
+        str(ext or "").strip().lstrip(".").lower(), "")
+
+
+def _img_srcs(steps_html):
+    """steps 富文本里的 <img src> 候选（含 data: URI）。"""
+    return re.findall(r'(?is)<img\b[^>]*?\bsrc\s*=\s*["\']([^"\']+)["\']',
+                      str(steps_html or ""))
+
+
+def _bug_mentions_images(bug):
+    """描述是否指向图片证据：steps 带 <img>、文字提「如图/截图…」、或附件含图片。"""
+    html = str(bug.get("steps") or "")
+    if "<img" in html.lower():
+        return True
+    text = str(bug.get("title") or "") + "\n" + _strip_html(html)
+    if _IMG_REF_RE.search(text):
+        return True
+    files = bug.get("files")
+    vals = files.values() if isinstance(files, dict) else (files if isinstance(files, list) else [])
+    for e in vals:
+        if isinstance(e, dict) and _img_ext_mime(
+                e.get("extension") or str(e.get("title") or "").rsplit(".", 1)[-1]):
+            return True
+    return False
+
+
+def _decode_data_uri(src):
+    m = re.match(r"(?is)^data:(image/[\w.+-]+);base64,(.+)$", src.strip())
+    if not m:
+        return "", b""
+    try:
+        raw = base64.b64decode(m.group(2))
+    except Exception:
+        return "", b""
+    mime = _img_mime(raw) or m.group(1).lower()
+    return mime, raw
+
+
+def _download_same_origin(src, cfg):
+    """下载禅道站内图片字节：仅同源（host:port 与配置的禅道一致），带会话多形态
+    尝试（zentaosid 查询参——新老两通道都认；再裸 GET 兜底），魔数校验。
+    非 200 / 非图 / 网络失败返回 b""。"""
+    try:
+        base = _raw_base(cfg.get("base_url"))
+    except ZenError:
+        return b""
+    url = src if "://" in str(src) else base + "/" + str(src).lstrip("/")
+    try:
+        u, bu = urllib.parse.urlsplit(url), urllib.parse.urlsplit(base)
+    except ValueError:
+        return b""
+    if (u.scheme or bu.scheme, u.netloc) != (bu.scheme, bu.netloc):
+        return b""          # 第三方图床一律不取（SSRF 边界外只放行同源）
+    url = _guard_url("%s://%s%s" % (bu.scheme, bu.netloc, u.path or "/")
+                     + (("?" + u.query) if u.query else ""))
+    tries = []
+    for sid in (str(_OLD.get("sid") or ""), str(_TOKEN.get("v") or "")):
+        if sid:
+            tries.append(url + ("&" if "?" in url else "?") + "zentaosid=" + sid)
+    tries.append(url)
+    for t in tries:
+        try:
+            req = urllib.request.Request(t, headers={"Accept": "*/*"})
+            with _no_redirect_opener().open(req, timeout=HTTP_TIMEOUT) as r:
+                if r.status != 200:
+                    continue
+                raw = r.read(20 * 1024 * 1024)
+        except Exception:
+            continue
+        if _img_mime(raw):
+            return raw
+    return b""
+
+
+def _file_entry_urls(fid, entry):
+    """详情 files 条目 → 下载候选 URL（entry 自带地址优先，再猜经典路由）。"""
+    out = []
+    if isinstance(entry, dict):
+        for k in ("download", "url", "href"):
+            v = str(entry.get(k) or "").strip()
+            if v.startswith(("http://", "https://", "/")):
+                out.append(v)
+        ext = _img_ext_mime(entry.get("extension")
+                            or str(entry.get("title") or "").rsplit(".", 1)[-1])
+    else:
+        ext = ""
+    if ext:
+        ext = ext.split("/", 1)[1]
+        for route in ("file-read-%s.%s", "file-download-%s.%s"):
+            out.append("/" + route % (fid, ext))
+    return out
+
+
+def _prep_image_pairs(pairs):
+    """[(mime, 原始字节)] → [(mime, b64)]。缩放规则与内置智能体同款（长边 1568、
+    超大转 JPEG q85、单张 5MB 剔除）；就地实现不跨层 import（架构 L2→L1 反向禁）。"""
+    try:
+        from PIL import Image
+    except Exception:
+        Image = None
+    out = []
+    for mime, raw in pairs or []:
+        data = raw
+        if Image is not None:
+            try:
+                im = Image.open(io.BytesIO(raw))
+                im.load()
+                w, h = im.size
+                edge = max(w, h, 1)
+                if edge > 1568:
+                    r = 1568 / float(edge)
+                    im = im.resize((max(1, round(w * r)), max(1, round(h * r))),
+                                   Image.LANCZOS)
+                fmt = {"image/png": "PNG", "image/jpeg": "JPEG",
+                       "image/gif": "GIF", "image/webp": "WEBP"}.get(mime, "PNG")
+                buf = io.BytesIO()
+                im.save(buf, format=fmt)
+                data = buf.getvalue()
+            except Exception:
+                data = raw                # 解码失败退回原始字节
+            if mime != "image/jpeg" and len(data) > 3500 * 1024:
+                try:                      # 仍过大：转 JPEG q85（RGB 拍平 alpha）
+                    im = Image.open(io.BytesIO(data)).convert("RGB")
+                    buf = io.BytesIO()
+                    im.save(buf, format="JPEG", quality=85)
+                    data, mime = buf.getvalue(), "image/jpeg"
+                except Exception:
+                    pass
+        if len(data) > 5 * 1024 * 1024:
+            continue
+        out.append((mime, base64.b64encode(data).decode("ascii")))
+    return out
+
+
+def _bug_images(bug_id, steps_html, cfg=None):
+    """bug 关联图片 → [(mime, b64)]：steps 内嵌 <img>（含 data: URI）优先，
+    读不到再看详情 files 附件。下载带禅道会话、魔数校验、缩放管线同内置智能体。
+    拿不到返回 []（也缓存，15 分钟内不重复下载）。"""
+    key = str(bug_id or "")
+    now = time.time()
+    hit = _IMG_CACHE.get(key)
+    if hit and now - hit[0] < IMG_CACHE_TTL:
+        return hit[1]
+    cfg = cfg or _cfg()
+    pairs = []
+    for src in _img_srcs(steps_html):
+        s = src.strip()
+        if s.lower().startswith("data:"):
+            mime, raw = _decode_data_uri(s)
+            if raw:
+                pairs.append((mime or "image/png", raw))
+        elif not s.lower().startswith("javascript:"):
+            raw = _download_same_origin(s, cfg)
+            if raw:
+                pairs.append((_img_mime(raw), raw))
+    if not pairs and key:
+        try:
+            detail = _call("GET", "/bugs/%s" % key, cfg=cfg)
+            files = detail.get("files") if isinstance(detail, dict) else None
+            entries = (list(files.items()) if isinstance(files, dict)
+                       else [(str(i), e) for i, e in enumerate(files or [])])
+            for fid, entry in entries:
+                mime = _img_ext_mime(
+                    (entry.get("extension") if isinstance(entry, dict) else "")
+                    or str(entry.get("title") if isinstance(entry, dict) else "").rsplit(".", 1)[-1])
+                if not mime:
+                    continue          # 日志/压缩包等非图附件不取
+                for u in _file_entry_urls(fid, entry):
+                    raw = _download_same_origin(u, cfg)
+                    if raw:
+                        pairs.append((_img_mime(raw), raw))
+                        break
+        except Exception:
+            log.debug("zentao: bug 附件图片获取失败", exc_info=True)
+    imgs = _prep_image_pairs(pairs[:TRIAGE_MAX_IMAGES])
+    if len(_IMG_CACHE) > 128:
+        for k in sorted(_IMG_CACHE, key=lambda k: _IMG_CACHE[k][0])[:64]:
+            _IMG_CACHE.pop(k, None)
+    _IMG_CACHE[key] = (now, imgs)
+    return imgs
+
+
+def _save_bug_attachments(workdir, bug_id, imgs):
+    """bug 截图落工作目录 _attachments/（create_task 的 dict 清单形态直用）。
+    返回 [{name,size,mime,path}]；失败返回 []。"""
+    if not imgs:
+        return []
+    try:
+        adir = Path(workdir) / "_attachments"
+        adir.mkdir(parents=True, exist_ok=True)
+        ext = {"image/png": ".png", "image/jpeg": ".jpg", "image/gif": ".gif",
+               "image/webp": ".webp"}
+        items = []
+        for i, (mime, b64) in enumerate(imgs, 1):
+            name = "zenbug%s-%d%s" % (bug_id, i, ext.get(mime, ".png"))
+            data = base64.b64decode(b64)
+            (adir / name).write_bytes(data)
+            items.append({"name": name, "size": len(data), "mime": mime,
+                          "path": "_attachments/" + name})
+        return items
+    except Exception:
+        log.debug("zentao: bug 截图落附件失败", exc_info=True)
+        return []
+
+
+def _claimable(bug, profile):
+    """认领过滤：active + 指派 + 严重度（产品由档案本身界定）。"""
+    if str(bug.get("status") or "") != "active":
+        return False
+    assigned = str(profile.get("assigned_to") or "").strip()
+    cap = int(profile.get("severity_cap") or 0)
+    if assigned and _acct(bug.get("assignedTo")) != assigned:
+        return False
+    if cap and _severity(bug) > cap:
+        return False
+    return True
+
+
+def list_bugs(cfg, product_id, profile=None):
+    """拉一个产品下的激活 bug（分页，总量封顶 MAX_BUGS）。
+
+    REST v1 支持时把状态/指派人过滤下推给禅道，减少无关 bug 的分页传输；
+    本地的 _claimable 仍保留，兼容老接口或禅道忽略未知查询参数的情况。
+    """
+    out = []
+    for page in range(1, 6):
+        query = {"page": page, "limit": PAGE_LIMIT, "status": "active"}
+        assigned = str((profile or {}).get("assigned_to") or "").strip()
+        if assigned:
+            query["assignedTo"] = assigned
+        d = _call("GET", "/products/%s/bugs?%s" % (
+            product_id, urllib.parse.urlencode(query)), cfg=cfg)
+        bugs = d.get("bugs") if isinstance(d, dict) else None
+        if not isinstance(bugs, list):
+            raise ZenError("禅道 bug 列表响应形状不对（预期 bugs 数组）")
+        out.extend(b for b in bugs if isinstance(b, dict))
+        total = 0
+        try:
+            total = int(d.get("total") or 0)
+        except (TypeError, ValueError):
+            pass
+        if len(out) >= MAX_BUGS or (total and len(out) >= total) or len(bugs) < PAGE_LIMIT:
+            break
+    return out[:MAX_BUGS]
+
+
+def weekly_brief(limit=12):
+    """本周工作素材（工作汇报起草注入用，2026-09-23 Weekly Report Generator 借鉴
+    ——「从工作系统取数入素材」）：分配给我的活跃任务 + 指派给我的活跃 Bug 摘要。
+    尽力而为：未配置/不可达/响应形状不对一律返回空串，绝不阻塞汇报起草。
+    多产品档案逐个聚合（同 _scan 口径），条目总量封顶 limit*2。"""
+    lines = []
+    try:
+        cfg = _cfg()
+        for profile in _profiles(cfg):
+            assigned = str(profile.get("assigned_to") or "").strip()
+            if not assigned:
+                continue
+            try:
+                d = _call("GET", "/tasks?%s" % urllib.parse.urlencode(
+                    {"assignedTo": assigned, "limit": limit}), cfg=cfg)
+                tasks = d.get("tasks") if isinstance(d, dict) else None
+                for t in (tasks or [])[:limit]:
+                    if isinstance(t, dict) and t.get("status") in ("wait", "doing"):
+                        lines.append("- [任务#%s] %s（%s）" % (
+                            t.get("id"), t.get("name") or "", t.get("status") or ""))
+            except Exception:
+                pass
+            try:
+                for b in list_bugs(cfg, profile.get("product"), profile)[:limit]:
+                    lines.append("- [Bug#%s] %s（严重级 %s）" % (
+                        b.get("id"), b.get("title") or "",
+                        b.get("severity") or b.get("status") or ""))
+            except Exception:
+                pass
+    except Exception:
+        return ""
+    return "\n".join(lines[:limit * 2])
+
+
+def fetch_modules(product_id):
+    """拉产品模块清单（模块路由配置辅助）。接口不存在/失败报人话，提示手填 ID。"""
+    _ensure_loaded()
+    cfg = _cfg()
+    try:
+        d = _call("GET", "/products/%s/modules" % product_id, cfg=cfg)
+    except ZenError as e:
+        msg = str(e)
+        # 只有「接口不存在」类错误才补「没有该接口」提示；老通道自己报的
+        # 空树/脚本回包已带人话结论，别再叠一层误导
+        if "404" in msg or "不认识" in msg:
+            msg += "——也可能你的禅道没有该接口：请在禅道产品视图 URL 里查模块 ID 手工填写"
+        return {"ok": False, "error": msg}
+    # 形状兼容器（{_list} / {modules:[...]} / {id:{...}} 字典 / 裸数组）
+    items = _list_items(d, "modules")
+    if items is None:
+        items = d.get("_list") if isinstance(d.get("_list"), list) else None
+    out = []
+    for m in items or []:
+        if isinstance(m, dict) and m.get("id"):
+            out.append({"id": m.get("id"), "name": str(m.get("name") or "")})
+    if not out:
+        if isinstance(d, dict) and set(d.keys()) == {"_list"}:
+            # 自家归一包装（裸数组进 _api/_old_route 时打的包）：空清单=
+            # 产品真没模块，不是形状问题，别拿顶层键吓人
+            return {"ok": False,
+                    "error": "禅道回了空模块清单（该产品可能没建模块）——请手工填模块 ID"}
+        # 带响应形状摘要帮排查（顶层键名，不吐正文——用户实测反馈
+        # 「响应形状不认识」却看不到真实形状，没法报修）
+        keys = ",".join(sorted(str(k) for k in d.keys())) if isinstance(d, dict) else type(d).__name__
+        return {"ok": False,
+                "error": "模块清单为空或响应形状不认识（响应顶层键：%s）——请手工填模块 ID" % (keys or "非字典")}
+    return {"ok": True, "modules": out[:200]}
+
+
+def _list_items(d, key):
+    """禅道 REST v1 清单响应形状兼容：{key:[...]} / 裸数组(_list) / {id:{...}} 字典。"""
+    if not isinstance(d, dict):
+        return None
+    if isinstance(d.get(key), list):
+        return d[key]
+    if isinstance(d.get("_list"), list):
+        return d["_list"]
+    vals = [v for v in d.values() if isinstance(v, dict) and v.get("id")]
+    return vals or None
+
+
+def fetch_products():
+    """拉产品清单（产品 ID 下拉选择辅助）。接口不存在/失败返回 ok:False+人话。"""
+    _ensure_loaded()
+    cfg = _cfg()
+    out, total = [], 0
+    try:
+        page = 1
+        while page <= 3:
+            d = _call("GET", "/products?page=%d&limit=100" % page, cfg=cfg)
+            items = _list_items(d, "products")
+            if not items:
+                break
+            out.extend(it for it in items
+                       if isinstance(it, dict) and str(it.get("id") or "") != "")
+            try:
+                total = int(d.get("total") or 0)
+            except (TypeError, ValueError, AttributeError):
+                total = 0
+            if not total or len(out) >= total:
+                break
+            page += 1
+    except ZenError as e:
+        return {"ok": False, "error": "%s——请先「测试连接」确认地址与账号可用" % e}
+    if not out:
+        return {"ok": False, "error": "禅道里没有产品，或响应形状不认识（REST API 需 ≥15）"}
+    seen, uniq = set(), []
+    for m in out:
+        try:
+            pid = int(m["id"])
+        except (TypeError, ValueError):
+            pid = m["id"]
+        if pid in seen:
+            continue
+        seen.add(pid)
+        uniq.append({"id": pid, "name": str(m.get("name") or ""),
+                     "status": str(m.get("status") or "")})
+    uniq.sort(key=lambda x: str(x["id"]))
+    return {"ok": True, "products": uniq[:500]}
+
+
+def fetch_users():
+    """拉禅道账号清单（负责人/转派下拉辅助）。失败返回 ok:False，前端仍可手填。"""
+    _ensure_loaded()
+    cfg = _cfg()
+    try:
+        d = _call("GET", "/users?limit=1000", cfg=cfg)
+    except ZenError as e:
+        return {"ok": False, "error": "%s——账号清单拉不到，仍可手填账号" % e}
+    items = None
+    if isinstance(d, dict):
+        if isinstance(d.get("users"), list):
+            items = d["users"]
+        elif isinstance(d.get("_list"), list):
+            items = d["_list"]
+        else:
+            items = []
+            for k, v in d.items():      # {账号: 用户} 字典形状
+                if not isinstance(v, dict):
+                    continue
+                u = dict(v)
+                u.setdefault("account", str(k))
+                items.append(u)
+    out, seen = [], set()
+    for u in items or []:
+        if not isinstance(u, dict):
+            continue
+        acct = str(u.get("account") or "").strip()
+        if not acct or acct in seen:
+            continue
+        seen.add(acct)
+        out.append({"account": acct, "realname": str(u.get("realname") or "")})
+    if not out:
+        return {"ok": False, "error": "账号清单为空或响应形状不认识——仍可手填账号"}
+    out.sort(key=lambda x: x["account"])
+    return {"ok": True, "users": out[:500]}
+
+
+def test_connection(base_url=None, account=None, password=None):
+    """连接测试：探明通道登录 + 拉第一个产品的 bug 列表（有档案时）。返回 (ok, 人话结果)。"""
+    c = _cfg()
+    base_url = str(base_url if base_url is not None else c.get("base_url") or "").strip()
+    account = str(account if account is not None else c.get("account") or "").strip()
+    password = str(password if password is not None else c.get("password") or "").strip()
+    if not (base_url and account and password):
+        return False, "地址、账号、密码都要填全"
+    b = base_url.strip().rstrip("/")
+    probe = {"base_url": b, "account": account, "password": password}
+    try:
+        tok = _rest_login(b, account, password)
+    except ZenError as e:
+        return False, str(e)
+    if tok:
+        _MODE[b] = "rest"
+    else:
+        try:
+            _old_login(probe, _REST_LAST_ERR.get("msg", ""))
+            _MODE[b] = "old"
+        except ZenError as e:
+            return False, str(e)
+    tag = "老版 JSON 接口" if _MODE.get(b) == "old" else "REST v1"
+    try:
+        profiles = _profiles(c)
+        if profiles:
+            bugs = list_bugs(probe, profiles[0]["product"], profiles[0])
+            return True, "连接成功（%s），产品 %s 可访问（当前 %d 条 bug 在列表里）" % (
+                tag, profiles[0]["product"], len(bugs))
+    except ZenError as e:
+        return False, "登录成功（%s），但拉 bug 列表失败：%s" % (tag, e)
+    return True, "连接成功（%s；未配产品档案，跳过列表探测）" % tag
+
+
+# ---------------------------------------------------------------- 排查（triage）
+
+def _ai_triage(bug, profile):
+    """AI 兜底排查：单次 LLM 调用判端。不可用/解析失败返回 None（bookmeta 同款配方）。
+
+    图片守卫（#27754 案）：描述指向截图而图片读不到 → unknown 留人工，绝不拿
+    纯文本盲判。排查模型不会看图时也不甩人工——自动全库找看图候选（声明
+    image_in 的优先，名字推断兜底），候选逐个试、实测吃图成功回写能力标记。
+    """
+    try:
+        from . import modelhub, runner
+        images = []
+        if _bug_mentions_images(bug):
+            try:
+                images = _bug_images(bug.get("id"), bug.get("steps"))
+            except Exception:
+                log.debug("zentao: bug 图片获取失败", exc_info=True)
+            if not images:
+                return {"side": "unknown",
+                        "reason": "描述指向截图/图片但图片未能读取，证据不足，"
+                                  "留人工确认后定责",
+                        "imgs": 0}
+        orch = modelhub.resolve_orchestrator()
+        prov, model = orch if orch else (None, None)
+        repos = profile.get("repos") or {}
+        hints = profile.get("repo_hints") or {}
+
+        def _repo_desc(side):
+            hint = str(hints.get(side) or "").strip()
+            if hint:
+                return hint
+            wd = str((repos.get(side) or {}).get("workdir") or "").strip()
+            if wd:
+                return "目录 " + Path(wd).name
+            return "（未配置）"
+
+        lines = ["你是缺陷分诊员：根据缺陷描述判断问题属于哪个仓库端。", "",
+                 "【缺陷】#%s %s" % (bug.get("id"), str(bug.get("title") or ""))]
+        steps = _strip_html(bug.get("steps"))
+        if steps:
+            lines.append(steps[:3000])
+        if images:
+            lines.append("【截图】随本消息附 %d 张 bug 截图，判定前先看图，"
+                         "截图与文字冲突时以截图为准。" % len(images))
+        lines.append("")
+        lines.append("【仓库背景】后端仓库：%s；前端仓库：%s" % (_repo_desc("backend"),
+                                                        _repo_desc("frontend")))
+        lines.append("")
+        lines.append('只输出 JSON（不要别的文字）：{"side": "backend|frontend|both|not_ours", '
+                     '"reason": "一句话依据"}')
+        lines.append("判定口径：backend=纯后端问题；frontend=纯前端问题；both=两端都要改；"
+                     "not_ours=与这两个仓库无关（第三方服务/环境/需求变更/数据问题等）。")
+        prompt = "\n".join(lines)
+        if images:
+            cand = modelhub.vision_candidates(
+                prefer_provider_id=(prov or {}).get("id") or "",
+                prefer_model=model or "")
+            if not cand:
+                return {"side": "unknown",
+                        "reason": "所有已启用模型都不支持图片输入，无法看图定责"
+                                  "（请在模型管理里启用一个视觉模型）",
+                        "imgs": 0}
+            res, last_err = None, ""
+            for p, m, inferred in cand:
+                r = modelhub.chat(p["id"], m, prompt, max_tokens=500, timeout=90,
+                                  images=images)
+                if r.get("ok"):
+                    res = r
+                    if inferred:
+                        # 名字推断的候选实测吃图成功：回写能力标记（幂等），
+                        # 模型管理页图徽章同步可见，后续调用直接命中
+                        err = modelhub.set_model_caps(p["id"], m, True)
+                        if err:
+                            log.debug("zentao: 自动标记 image_in 失败：%s", err)
+                    break
+                last_err = str(r.get("error") or "")
+                log.warning("zentao: 视觉排查候选 %s/%s 不可用：%s",
+                            p.get("id"), m, last_err[:160])
+            if not res:
+                return {"side": "unknown",
+                        "reason": "看图排查全部候选失败（%s）" % last_err[:160],
+                        "imgs": 0}
+        else:
+            if not orch:
+                return None
+            res = modelhub.chat(prov["id"], model, prompt, max_tokens=500, timeout=90)
+            if not res.get("ok"):
+                log.warning("zentao: AI 排查失败：%s", res.get("error"))
+                return None
+        data = runner.extract_json(res.get("text") or "")
+        side = str((data or {}).get("side") or "").strip().lower()
+        if side not in TRIAGE_SIDES:
+            return None
+        return {"side": side, "reason": str((data or {}).get("reason") or "")[:300],
+                "imgs": len(images)}
+    except Exception:
+        log.debug("zentao: AI 排查异常", exc_info=True)
+        return None
+
+
+def _triage(bug, profile, cfg):
+    """排查定责：模块路由 > AI > unknown。返回 {side, reason, by, account}。"""
+    mid = str(bug.get("module") or "")
+    if mid:
+        for r in profile.get("module_routes") or []:
+            try:
+                if int(r.get("module") or 0) != int(mid):
+                    continue
+            except (TypeError, ValueError):
+                continue
+            side = str(r.get("side") or "")
+            if side in TRIAGE_SIDES:
+                return {"side": side, "reason": "模块 #%s 路由规则" % mid,
+                        "by": "rule", "account": str(r.get("account") or "").strip()}
+    if cfg.get("triage_ai"):
+        res = _ai_triage(bug, profile)
+        if res:
+            res["by"] = "ai"
+            res["account"] = ""
+            return res
+    return {"side": "unknown", "reason": "模块未命中路由且 AI 排查不可用", "by": "fallback",
+            "account": ""}
+
+
+# ---------------------------------------------------------------- 修复任务拉起
+
+def _repo_of(profile, side):
+    return (profile.get("repos") or {}).get(side) or dict(_REPO_DEFAULTS)
+
+
+def _goal_text(bug, side=None, images=0):
+    """bug → 修复目标提示词。side 给出时附端约束；images>0 时提示先读附件截图。"""
+    bid = bug.get("id")
+    lines = ["修复禅道 Bug #%s：%s" % (bid, str(bug.get("title") or "").strip())]
+    steps = _strip_html(bug.get("steps"))
+    if steps:
+        lines.append("")
+        lines.append("【重现步骤/问题描述】")
+        lines.append(steps[:4000])
+    sev, pri = _severity(bug), str(bug.get("pri") or "").strip()
+    meta = []
+    if sev:
+        meta.append("严重度 %s（1 最严重）" % sev)
+    if pri:
+        meta.append("优先级 %s" % pri)
+    if str(bug.get("module") or "").strip():
+        meta.append("模块 #%s" % bug["module"])
+    env = " / ".join(x for x in (str(bug.get("os") or "").strip(),
+                                 str(bug.get("browser") or "").strip()) if x)
+    if env:
+        meta.append("环境 " + env)
+    kw = str(bug.get("keywords") or "").strip()
+    if kw:
+        meta.append("关键字 " + kw)
+    if meta:
+        lines.append("")
+        lines.append("【元信息】" + "；".join(meta))
+    if images:
+        lines.append("")
+        lines.append("【截图】bug 截图共 %d 张，已放在工作目录 _attachments/（文件名 "
+                     "zenbug*）；先读图再动手——需求以截图为准，不要凭文字想象。"
+                     % images)
+    lines.append("")
+    lines.append("【要求】只修这个 bug，不做无关重构；改动最小化；修完自查不引入回归。")
+    if side in SIDES:
+        lines.append("【端约束】本任务只负责【%s】部分；%s部分由别人处理，不要越界改动。"
+                     % (SIDE_CN[side],
+                        SIDE_CN["frontend" if side == "backend" else "backend"]))
+    lines.append("完成后给出修改说明。")
+    return "\n".join(lines)
+
+
+def _launch_fix(bug, profile, side, cfg):
+    """为一个 bug 的某一端建修复任务并立即启动。"""
+    bid = bug.get("id")
+    repo = _repo_of(profile, side)
+    wd = str(repo.get("workdir") or "").strip() or settings.default_workdir()
+    title = ("[禅道#%s][%s] %s" % (bid, SIDE_CN.get(side, side),
+                                   str(bug.get("title") or "").strip())).strip()[:60]
+    # bug 截图随任务走：修复智能体跟排查同样需要看图，否则修出来的就是想象中的 bug
+    imgs = []
+    try:
+        if _bug_mentions_images(bug):
+            imgs = _bug_images(bid, bug.get("steps"))
+    except Exception:
+        imgs = []
+    atts = _save_bug_attachments(str(wd), bid, imgs)
+    payload = {"type": "code", "title": title, "goal": _goal_text(bug, side, images=len(atts)),
+               "workdir": wd}
+    if atts:
+        payload["attachments"] = atts
+    base = str(repo.get("git_rev") or "").strip()
+    if not base:
+        # 档案没配基线就取落单时点的 HEAD：让修复走任务分支隔离——修完自动
+        # 提交在任务分支上，对账合并才真正落库，resolve 才站得住（#27697 案：
+        # 无基线任务修完只躺在工作区，闭环永远不会替你提交）。
+        base = _head_rev(wd)
+    if base:
+        payload["git_rev"] = base
+    if str(repo.get("verify_command") or "").strip():
+        payload["verify_command"] = str(repo["verify_command"]).strip()
+    task = None
+    run = None
+    try:
+        task = store.create_task(payload)
+        run = store.create_run("orchestration", task["title"], task_id=task["id"])
+        store.update_task_status(task["id"], "queued")
+        jobs.enqueue({"kind": "orchestration", "run_id": run["id"],
+                      "task_id": task["id"]})
+        return task, run
+    except Exception:
+        log.exception("zentao: 修复任务启动失败 bug=%s side=%s task=%s run=%s",
+                      bid, side, (task or {}).get("id"), (run or {}).get("id"))
+        if run:
+            try:
+                store.update_run(run["id"], status="failed",
+                                 error="禅道修复任务启动失败，本次未排队，请稍后重试",
+                                 ended_at=time.strftime("%Y-%m-%d %H:%M:%S"))
+            except Exception:
+                log.exception("zentao: 修复运行失败收口失败 run=%s", run.get("id"))
+        if task:
+            try:
+                store.update_task_status(task["id"], "failed")
+            except Exception:
+                log.exception("zentao: 修复任务失败收口失败 task=%s", task.get("id"))
+        raise
+
+
+# ---------------------------------------------------------------- 回写文本
+
+def _diffstat(workdir, task_id):
+    """任务分支相对基线的改动统计（人话一行）。拿不到返回空串。"""
+    try:
+        from . import gitmod, runner
+        if task_id and workdir:
+            br = gitmod.branch_name(task_id)
+            r = runner.run_process(
+                argv=["git", "-C", workdir, "diff", "--shortstat", "HEAD..." + br],
+                timeout=20)
+            if r.get("ok"):
+                return (r.get("stdout") or "").strip()
+    except Exception:
+        log.debug("zentao: diffstat 失败", exc_info=True)
+    return ""
+
+
+def _fix_summary(claim, run, profile, side):
+    """我方某一端的修复摘要（转派评论与 resolve 报告共用的事实部分）。"""
+    lines = []
+    git = (run or {}).get("git") or {}
+    commit = str(git.get("commit") or "").strip()
+    from_branch = str(git.get("from_branch") or "").strip()
+    tid = _task_of(claim, side).get("task_id") or ""
+    if commit:
+        lines.append("【%s】修复提交 %s%s" % (SIDE_CN.get(side, side), commit,
+                     "（已合并回 %s）" % from_branch if from_branch else ""))
+    stat = _diffstat(str(_repo_of(profile, side).get("workdir") or ""), tid)
+    if stat:
+        lines.append("【%s】改动统计：%s" % (SIDE_CN.get(side, side), stat))
+    v = (run or {}).get("verdict") or {}
+    if v:
+        lines.append("【%s】验证结论：%s" % (SIDE_CN.get(side, side),
+                     "通过（评审达标）" if v.get("pass") or v.get("publishable") else "完成"))
+    return lines
+
+
+def _task_of(claim, side):
+    for t in claim.get("tasks") or []:
+        if t.get("side") == side:
+            return t
+    return {}
+
+
+def _report_text(claim, runs, profile, cfg):
+    """resolve 评论：确定性事实（逐端汇总）。"""
+    tri = claim.get("triage") or {}
+    lines = ["【CodeBee 自动修复报告】",
+             "Bug：#%s %s" % (claim.get("bug_id"), claim.get("title") or "")]
+    if tri.get("side"):
+        lines.append("排查结论：%s问题（%s）" % (tri.get("side"),
+                                               tri.get("reason") or tri.get("by") or "按规则"))
+    for t in claim.get("tasks") or []:
+        lines.extend(_fix_summary(claim, runs.get(t.get("run_id")), profile, t.get("side")))
+    lines.append("（本条由 CodeBee 禅道集成自动回写）")
+    return "\n".join(lines)
+
+
+def _transfer_text(claim, profile, fixed_runs, target_side):
+    """转派评论：排查结论 + 我方已做工作 + 请对方继续。"""
+    tri = claim.get("triage") or {}
+    lines = ["【CodeBee 排查转派】",
+             "Bug：#%s %s" % (claim.get("bug_id"), claim.get("title") or ""),
+             "排查结论：%s问题——%s" % (SIDE_CN.get(target_side, target_side),
+                                      tri.get("reason") or tri.get("by") or "按规则")]
+    if tri.get("imgs"):
+        lines.append("（判定依据含 %d 张 bug 截图）" % tri["imgs"])
+    for side, run in (fixed_runs or []):
+        lines.extend(_fix_summary(claim, run, profile, side))
+    lines.append("请%s负责人接手处理；本 bug 保持激活，处理完请按正常流程解决。"
+                 % SIDE_CN.get(target_side, target_side))
+    lines.append("（本条由 CodeBee 禅道集成自动回写）")
+    return "\n".join(lines)
+
+
+def _fail_text(claim, failed_tasks, runs, internal=False):
+    tri = claim.get("triage") or {}
+    lines = ["【CodeBee 自动修复未成功】",
+             "Bug：#%s %s" % (claim.get("bug_id"), claim.get("title") or "")]
+    if tri.get("side"):
+        lines.append("排查结论：%s（%s）" % (tri.get("side"), tri.get("reason") or ""))
+    for t in failed_tasks:
+        r = runs.get(t.get("run_id")) or {}
+        why = str(r.get("error") or "").strip() or ("运行状态 " + str(r.get("status") or ""))
+        lines.append("【%s】失败：%s" % (SIDE_CN.get(t.get("side"), t.get("side")), why[:300]))
+    if internal:
+        lines.append("（失败原因是 CodeBee 工具自身异常，不代表修复结论；"
+                     "请人工排查 CodeBee 或续跑修复任务）")
+    lines.append("CodeBee 修复任务：%s（可人工续跑或接管）；本 bug 保持待处理。"
+                 % "、".join(t.get("task_id") or "?" for t in claim.get("tasks") or []))
+    return "\n".join(lines)
+
+
+# ---------------------------------------------------------------- 禅道写回动作
+
+def _bug_opened_by(cfg, bug_id):
+    try:
+        d = _call("GET", "/bugs/%s" % bug_id, cfg=cfg)
+        return _acct(d.get("openedBy"))
+    except ZenError:
+        return ""
+
+
+def _ensure_resolved(cfg, bug_id, comment, assign_to):
+    """resolve（幂等）：已是 resolved/closed 视为成功。"""
+    cur = _call("GET", "/bugs/%s" % bug_id, cfg=cfg)
+    if str(cur.get("status") or "") in ("resolved", "closed"):
+        return True
+    body = {"resolution": "fixed", "resolvedBuild": "trunk", "comment": comment}
+    if assign_to:
+        body["assignedTo"] = assign_to
+    operation_id = operations.begin("zentao:bug:%s:resolve" % bug_id, body,
+                                    metadata={"bug_id": bug_id, "action": "resolve"})
+    try:
+        result = _call("POST", "/bugs/%s/resolve" % bug_id, cfg=cfg, body=body)
+        receipt = str((result or {}).get("id") or bug_id)
+        operations.confirm(operation_id, remote_receipt=receipt,
+                           metadata={"bug_id": bug_id, "action": "resolve"})
+    except Exception as exc:
+        operations.finish_exception(operation_id, exc)
+        raise
+    return True
+
+
+def _transfer(cfg, bug_id, target, comment):
+    """转派：PUT assignedTo+comment（bug 保持激活）。target 空=只评论。抛 ZenError。
+
+    禅道 PUT /bugs/{id} 不带 assignedTo 会把指派人清空（#27783 案：只评论
+    也出了「指派给(空)」记录），空 target 必须先 GET 带回当前指派人；GET
+    不到就整个不发——宁可不评论，不可盲写清指派。
+    """
+    body = {"comment": comment}
+    if target:
+        body["assignedTo"] = target
+    else:
+        cur = _call("GET", "/bugs/%s" % bug_id, cfg=cfg)
+        body["assignedTo"] = _acct(cur.get("assignedTo")) or ""
+    operation_id = operations.begin("zentao:bug:%s:transfer" % bug_id, body,
+                                    metadata={"bug_id": bug_id, "action": "transfer"})
+    try:
+        result = _call("PUT", "/bugs/%s" % bug_id, cfg=cfg, body=body)
+        receipt = str((result or {}).get("id") or bug_id)
+        operations.confirm(operation_id, remote_receipt=receipt,
+                           metadata={"bug_id": bug_id, "action": "transfer"})
+    except Exception as exc:
+        operations.finish_exception(operation_id, exc)
+        raise
+
+
+def _notify(text):
+    try:
+        from . import notify
+        notify.push_text(text)
+    except Exception:
+        log.debug("zentao: 群通知失败", exc_info=True)
+
+
+def _merge_branch(workdir, task):
+    try:
+        from . import gitmod
+        return gitmod.merge_task_branch(workdir, task)
+    except Exception as e:
+        return False, "合并异常：%s" % e, None
+
+
+def _head_rev(workdir):
+    """工作目录当前 HEAD 短哈希（作任务隔离基线）。非仓库/拿不到返回空串。"""
+    try:
+        from . import gitmod, runner
+        r = runner.run_process(
+            argv=["git", "-C", str(workdir), "rev-parse", "--short", "HEAD"], timeout=20)
+        if r.get("ok"):
+            rev = (r.get("stdout") or "").strip()
+            if rev and gitmod.valid_rev(rev):
+                return rev
+    except Exception:
+        log.debug("zentao: 读取 HEAD 失败", exc_info=True)
+    return ""
+
+
+def _uncommitted(workdir):
+    """工作目录里有没有未提交的已跟踪改动（untracked 不算——CLI 噪音文件太多）。
+
+    返回 (是否脏, 人话明细)。目录不是 git 仓库视为不脏（无提交语义，放行）。
+    """
+    try:
+        from . import runner
+        if not workdir:
+            return False, ""
+        r = runner.run_process(
+            argv=["git", "-C", str(workdir), "status", "--porcelain", "-uno", "--", "."],
+            timeout=20)
+        if r.get("ok"):
+            dirty = [ln.strip() for ln in (r.get("stdout") or "").splitlines() if ln.strip()]
+            if dirty:
+                return True, "未提交改动 %d 处（如 %s）" % (len(dirty), dirty[0][:80])
+    except Exception:
+        log.debug("zentao: 工作区状态检查失败", exc_info=True)
+    return False, ""
+
+
+def _set_claim(bid, **patch):
+    with _LOCK:
+        c = _STATE["claims"].get(str(bid))
+        if c is None:
+            return
+        c.update(patch)
+        c["at"] = time.strftime("%Y-%m-%d %H:%M:%S")
+        _save_locked()
+
+
+def _route_account(profile, tri, side):
+    """转派/升级目标：模块路由 account > 端负责人。"""
+    if tri and tri.get("account"):
+        return str(tri["account"]).strip()
+    return str((profile.get("owners") or {}).get(side) or "").strip()
+
+
+# ---------------------------------------------------------------- 对账回写
+
+def _finish_ok(claim, cfg):
+    """我方任务全部 done：合并 → （需要则）转派 → 否则 resolve。"""
+    bid = str(claim.get("bug_id"))
+    profile = _profile_for(cfg, claim.get("product")) or {}
+    tri = claim.get("triage") or {}
+    tasks = claim.get("tasks") or []
+    runs = {t.get("run_id"): store.get_run(t.get("run_id") or "") for t in tasks}
+    # 1) 逐任务合并（任一失败即停：不转派不 resolve，绝不带着没落库的修复转派）
+    if not cfg.get("auto_merge"):
+        _set_claim(bid, state="done_manual",
+                   note="修复完成；auto_merge 已关，请人工合并落库后到禅道解决 bug")
+        return
+    for t in tasks:
+        task = store.get_task(t.get("task_id") or "")
+        if task is None or not task.get("git_rev"):
+            continue
+        ok, err, _info = _merge_branch(str(_repo_of(profile, t.get("side")).get("workdir")
+                                           or settings.default_workdir()), task)
+        if not ok:
+            _set_claim(bid, state="merge_failed",
+                       note="【%s】任务分支合并失败：%s（不转派不 resolve，留人工）"
+                            % (SIDE_CN.get(t.get("side"), t.get("side")), err))
+            _notify("🐛❌ 禅道 Bug #%s 的%s修复代码合并失败：%s\n修复任务：%s"
+                    % (bid, SIDE_CN.get(t.get("side"), ""), err, t.get("task_id") or "?"))
+            return
+    # 1.5) 落库闸：没基线的任务不走任务分支隔离，改动只躺在工作区——带着
+    # 未提交的已跟踪改动点 resolve 是对禅道撒谎（#27697 案），拦下转人工。
+    for t in tasks:
+        task = store.get_task(t.get("task_id") or "")
+        if task is None or task.get("git_rev"):
+            continue
+        dirty, detail = _uncommitted(str(_repo_of(profile, t.get("side")).get("workdir")
+                                         or settings.default_workdir()))
+        if dirty:
+            _set_claim(bid, state="done_manual",
+                       note="修复完成但工作区有未提交改动（%s），已阻止自动 resolve"
+                            "——请人工提交后到禅道解决" % detail)
+            _notify("🐛⚠️ 禅道 Bug #%s 修复完成但未提交落库，已阻止自动 resolve\n%s"
+                    % (bid, claim.get("title") or ""))
+            return
+    # 2) 需要转派的端 = 判定端 - 我方端（both 且我方只管一端时非空）
+    verdict_sides = ([tri["side"]] if tri.get("side") in SIDES
+                     else list(SIDES) if tri.get("side") == "both" else [])
+    other_sides = [s for s in verdict_sides if s not in (profile.get("our_sides") or [])]
+    if other_sides:
+        targets = {}
+        for s in other_sides:
+            tgt = _route_account(profile, tri, s)
+            if not tgt:
+                _set_claim(bid, state="need_manual",
+                           note="我方已修完，但未配置%s负责人，无法转派——请人工转派" % SIDE_CN[s])
+                _notify("🐛⚠️ 禅道 Bug #%s 我方部分已修完，但未配置%s负责人，请人工转派"
+                        % (bid, SIDE_CN[s]))
+                return
+            targets[s] = tgt
+        fixed_runs = [(t.get("side"), runs.get(t.get("run_id"))) for t in tasks]
+        for s in other_sides:
+            try:
+                _transfer(cfg, bid, targets[s], _transfer_text(claim, profile, fixed_runs, s))
+            except ZenError as e:
+                attempts = int(claim.get("attempts") or 0) + 1
+                if attempts >= RESOLVE_MAX_ATTEMPTS:
+                    _set_claim(bid, attempts=attempts, state="resolve_failed",
+                               note="转派连续 %d 次失败：%s" % (attempts, e))
+                    _notify("🐛⚠️ 禅道 Bug #%s 修完但转派失败：%s" % (bid, e))
+                else:
+                    _set_claim(bid, attempts=attempts, state="fixing",
+                               note="转派第 %d 次失败，下轮重试：%s" % (attempts, e))
+                return
+        _set_claim(bid, state="transferred",
+                   note="我方（%s）已修完并合并，转派给 %s"
+                        % ("/".join(SIDE_CN.get(t.get("side"), "") for t in tasks),
+                           "、".join(SIDE_CN[s] + ":" + targets[s] for s in other_sides)))
+        _notify("🐛🔁 禅道 Bug #%s 我方已修完，转派 %s\n%s"
+                % (bid, "、".join(SIDE_CN[s] + ":" + targets[s] for s in other_sides),
+                   claim.get("title") or ""))
+        return
+    # 3) 全部我方端：resolve（幂等）+ 指回报告人
+    if not cfg.get("auto_resolve"):
+        _set_claim(bid, state="done_manual",
+                   note="修复完成；auto_resolve 已关，请人工确认后到禅道解决 bug")
+        return
+    report = _report_text(claim, runs, profile, cfg)
+    opened = _bug_opened_by(cfg, bid)
+    try:
+        _ensure_resolved(cfg, bid, report, opened)
+    except ZenError as e:
+        attempts = int(claim.get("attempts") or 0) + 1
+        if attempts >= RESOLVE_MAX_ATTEMPTS:
+            _set_claim(bid, attempts=attempts, state="resolve_failed",
+                       note="resolve 连续 %d 次失败：%s" % (attempts, e))
+            _notify("🐛⚠️ 禅道 Bug #%s 修复完成但回写禅道失败：%s" % (bid, e))
+        else:
+            _set_claim(bid, attempts=attempts, state="fixing",
+                       note="resolve 第 %d 次失败，下轮重试：%s" % (attempts, e))
+        return
+    _set_claim(bid, state="resolved", note="已 resolve(fixed)%s"
+               % ("，指回报告人 " + opened if opened else ""))
+    _notify("🐛✅ 禅道 Bug #%s 已修复并 resolve\n%s\n修复任务：%s"
+            % (bid, claim.get("title") or "", "、".join(t.get("task_id") or "" for t in tasks)))
+
+
+# CodeBee 自身崩溃形态：run.error 是裸 Python 异常 repr（如 NameError(...)，
+# 0.1.63 _TRANSIENT 缩进事故实测形态）或带 Python traceback。这不是「修不动
+# bug」——转派给端负责人等于拿自家工具故障甩锅。冒号形态（TimeoutError: x）
+# 多为 CLI/被测仓的真实输出，不算内部崩溃。
+_CRASH_REPR_RE = re.compile(r"^[A-Za-z_]\w*(?:Error|Exception|Interrupt)\(")
+
+
+def _looks_internal_crash(err):
+    e = str(err or "").strip()
+    if not e:
+        return False
+    return ("Traceback (most recent call last" in e
+            or _CRASH_REPR_RE.match(e) is not None)
+
+
+def _finish_failed(claim, cfg):
+    """我方任一任务失败：评论尝试记录 + 转派该端负责人（有配则转）。
+
+    CodeBee 自身崩溃（内部异常）不转派——工具故障不是「修不动」，转出去是
+    甩锅，只评论+群通知人工。负责人未配置同样只评论。
+    """
+    bid = str(claim.get("bug_id"))
+    profile = _profile_for(cfg, claim.get("product")) or {}
+    tasks = claim.get("tasks") or []
+    runs = {t.get("run_id"): store.get_run(t.get("run_id") or "") for t in tasks}
+    failed = [t for t in tasks
+              if str((runs.get(t.get("run_id")) or {}).get("status") or "") not in
+              ("queued", "running", "done")]
+    internal = any(_looks_internal_crash((runs.get(t.get("run_id")) or {}).get("error"))
+                   for t in failed)
+    text = _fail_text(claim, failed, runs, internal=internal)
+    # 升级目标：取第一个失败端的路由账号/负责人（内部崩溃不转派）
+    target = ""
+    if not internal:
+        for t in failed:
+            target = _route_account(profile, claim.get("triage") or {}, t.get("side"))
+            if target:
+                break
+    try:
+        _transfer(cfg, bid, target, text)
+        if internal:
+            note = "CodeBee 内部异常，已评论说明（不转派，请人工排查工具链）"
+            state = "commented"
+        elif target:
+            note = "已评论说明并转派 %s" % target
+            state = "escalated"
+        else:
+            side = str((failed[0].get("side") if failed else "") or "")
+            note = "已评论说明（%s负责人未配置，未转派）" % SIDE_CN.get(side, side or "该端")
+            state = "commented"
+    except ZenError as e:
+        log.warning("zentao: bug %s 失败评论未送达（群通知兜底）：%s", bid, e)
+        note = "修复失败，评论未送达：%s" % e
+        state = "commented"
+    _set_claim(bid, state=state, note=note)
+    why = ("CodeBee 内部异常，未转派" if internal
+           else ("已转派 " + target if target else "负责人未配置，未转派"))
+    _notify("🐛❌ 禅道 Bug #%s 自动修复未成功（%s）\n%s\n修复任务：%s"
+            % (bid, why,
+               claim.get("title") or "", "、".join(t.get("task_id") or "" for t in tasks)))
+
+
+def _refresh_claim_runs(claim):
+    """把 claim.tasks 的 run_id 对齐到各任务最新一次运行，返回 (tasks, 是否换绑)。
+
+    领单时钉住的 run 失败后，人工重试会换新 run——对账只认旧 run 会把已经
+    成功的修复判成失败（2026-09-22 #27697 误报失败评论案）。
+    """
+    fresh, changed = [], False
+    for t in claim.get("tasks") or []:
+        tid = str(t.get("task_id") or "")
+        if tid:
+            runs = store.task_runs(tid)
+            rid = str((runs[0].get("id") if runs else "") or "")
+            if rid and rid != t.get("run_id"):
+                t = dict(t, run_id=rid)
+                changed = True
+        fresh.append(t)
+    return fresh, changed
+
+
+def _reconcile(cfg):
+    """对账：fixing 中的 claim 查各 run 终态并回写。单条异常只跳过该条。"""
+    with _LOCK:
+        fixing = [dict(c) for c in _STATE["claims"].values() if c.get("state") == "fixing"]
+    for claim in fixing:
+        bid = str(claim.get("bug_id"))
+        try:
+            tasks, moved = _refresh_claim_runs(claim)
+            if moved:
+                _set_claim(bid, tasks=tasks)
+                claim = dict(claim, tasks=tasks)
+            runs = {t.get("run_id"): store.get_run(t.get("run_id") or "") for t in tasks}
+            if any(r is None for r in runs.values()):
+                _set_claim(bid, state="lost", note="运行记录不存在（可能被清理）")
+                continue
+            states = [str((r or {}).get("status") or "") for r in runs.values()]
+            if any(s in ("queued", "running") for s in states):
+                continue
+            if all(s == "done" for s in states):
+                _finish_ok(claim, cfg)
+            else:
+                _finish_failed(claim, cfg)
+        except Exception:
+            log.exception("zentao: claim %s 对账异常，跳过", bid)
+
+
+# ---------------------------------------------------------------- 扫描主流程
+
+def _sides_to_fix(profile, tri):
+    """判定端 ∩ 我方端。"""
+    tri_side = tri.get("side")
+    verdict_sides = [tri_side] if tri_side in SIDES else (list(SIDES) if tri_side == "both" else [])
+    return [s for s in verdict_sides if s in (profile.get("our_sides") or [])]
+
+
+def _route_one(bug, profile, cfg, notify=True):
+    """认领一个 bug：排查 → 分流（建任务/转派/留人工）。notify=False 用于
+    need_manual 存量的重排查（仍无解时不重复群通知）。返回动作文案或 None。"""
+    bid = str(bug.get("id") or "")
+    tri = _triage(bug, profile, cfg)
+    base_claim = {
+        "bug_id": bug.get("id"), "product": profile.get("product"),
+        "title": str(bug.get("title") or "")[:120],
+        "triage": tri, "tasks": [], "state": "fixing", "note": "",
+        "attempts": 0, "claimed_at": time.strftime("%Y-%m-%d %H:%M:%S"),
+    }
+    opened = _acct(bug.get("openedBy"))
+
+    if tri["side"] == "unknown":
+        base_claim["state"] = "need_manual"
+        base_claim["note"] = tri.get("reason") or "排查失败，留人工"
+        with _LOCK:
+            _STATE["claims"][bid] = base_claim
+            _save_locked()
+        if notify:
+            _notify("🐛❓ 禅道 Bug #%s 排查不出归属端（%s），留人工\n%s"
+                    % (bid, tri.get("reason") or "", bug.get("title") or ""))
+        return "need_manual"
+
+    if tri["side"] == "not_ours":
+        target = tri.get("account") or str(profile.get("owners").get("not_ours") or "") \
+            or opened
+        text = ("【CodeBee 排查转派】\nBug：#%s %s\n排查结论：非我方两个仓库的问题——%s\n"
+                "%s转回 %s 核实处理。\n（本条由 CodeBee 禅道集成自动回写）"
+                % (bid, bug.get("title") or "", tri.get("reason") or "按规则",
+                   ("（判定依据含 %d 张 bug 截图）\n" % tri["imgs"]) if tri.get("imgs") else "",
+                   target or "报告人"))
+        try:
+            _transfer(cfg, bid, target, text)
+            note = "非我方，已转派 %s" % (target or "（未指派，仅评论）")
+        except ZenError as e:
+            base_claim.update({"state": "need_manual", "note": "非我方转派失败：%s" % e})
+            with _LOCK:
+                _STATE["claims"][bid] = base_claim
+                _save_locked()
+            _notify("🐛⚠️ 禅道 Bug #%s 判定非我方但转派失败：%s" % (bid, e))
+            return "need_manual"
+        base_claim.update({"state": "transferred", "note": note})
+        with _LOCK:
+            _STATE["claims"][bid] = base_claim
+            _save_locked()
+        _notify("🐛↩️ 禅道 Bug #%s 判定非我方，已转派 %s\n%s"
+                % (bid, target or "报告人", bug.get("title") or ""))
+        return "transferred"
+
+    sides = _sides_to_fix(profile, tri)
+    if not sides:
+        # 纯对方端问题（含测试指错到我方账号的）：直接转派正确负责人
+        other = tri["side"] if tri["side"] in SIDES else None
+        if other is None:            # both 但我方两端都没配：整单转不了，留人工
+            base_claim["state"] = "need_manual"
+            base_claim["note"] = "双端问题但产品档案未配置我方端，请人工处理"
+            with _LOCK:
+                _STATE["claims"][bid] = base_claim
+                _save_locked()
+            _notify("🐛❓ 禅道 Bug #%s 为双端问题但未配我方端，留人工" % bid)
+            return "need_manual"
+        target = _route_account(profile, tri, other)
+        if not target:
+            base_claim["state"] = "need_manual"
+            base_claim["note"] = "判定为%s问题，但未配置%s负责人，无法转派" % (
+                SIDE_CN[other], SIDE_CN[other])
+            with _LOCK:
+                _STATE["claims"][bid] = base_claim
+                _save_locked()
+            _notify("🐛⚠️ 禅道 Bug #%s 判定为%s问题但未配负责人，请人工转派"
+                    % (bid, SIDE_CN[other]))
+            return "need_manual"
+        text = _transfer_text({"bug_id": bug.get("id"), "title": bug.get("title") or "",
+                               "triage": tri, "tasks": []}, profile, [], other)
+        try:
+            _transfer(cfg, bid, target, text)
+        except ZenError as e:
+            base_claim["state"] = "need_manual"
+            base_claim["note"] = "%s转派失败：%s" % (SIDE_CN[other], e)
+            with _LOCK:
+                _STATE["claims"][bid] = base_claim
+                _save_locked()
+            _notify("🐛⚠️ 禅道 Bug #%s 判定%s问题但转派失败：%s" % (bid, SIDE_CN[other], e))
+            return "need_manual"
+        base_claim.update({"state": "transferred",
+                           "note": "%s问题（测试原指派 %s），已转派 %s"
+                                   % (SIDE_CN[other], _acct(bug.get("assignedTo")) or "?", target)})
+        with _LOCK:
+            _STATE["claims"][bid] = base_claim
+            _save_locked()
+        _notify("🐛🔁 禅道 Bug #%s 判定%s问题，已转派 %s（原指派 %s）\n%s"
+                % (bid, SIDE_CN[other], target, _acct(bug.get("assignedTo")) or "?",
+                   bug.get("title") or ""))
+        return "transferred"
+
+    # 我方端：逐端建修复任务（任一端建不起来 → 整单留人工，不做半截修复）
+    tasks = []
+    try:
+        for side in sides:
+            task, run = _launch_fix(bug, profile, side, cfg)
+            tasks.append({"side": side, "task_id": task["id"], "run_id": run["id"],
+                          "state": "fixing"})
+    except Exception as e:
+        log.warning("zentao: bug %s 建修复任务失败：%s", bid, e)
+        base_claim["state"] = "need_manual"
+        base_claim["note"] = "建修复任务失败：%s" % e
+        with _LOCK:
+            _STATE["claims"][bid] = base_claim
+            _save_locked()
+        _notify("🐛⚠️ 禅道 Bug #%s 建修复任务失败：%s" % (bid, e))
+        return "need_manual"
+    base_claim["tasks"] = tasks
+    with _LOCK:
+        _STATE["claims"][bid] = base_claim
+        _save_locked()
+    _notify("🐛🔍 认领禅道 Bug #%s（%s问题·%s）：%s\n修复任务：%s"
+            % (bid, tri["side"], "由规则" if tri.get("by") == "rule" else "AI 判定",
+               bug.get("title") or "", "、".join(t["task_id"] for t in tasks)))
+    return "fixing"
+
+
+def _scan(cfg):
+    """按产品档案逐个拉 bug：认领新 bug + 重试 need_manual 的存量。返回认领数。"""
+    claimed = 0
+    profiles = _profiles(cfg)
+    if not profiles:
+        raise ZenError("未配置产品档案——请先在设置页添加产品并配置仓库")
+    for profile in profiles:
+        bugs = list_bugs(cfg, profile["product"], profile)
+        by_id = {str(b.get("id") or ""): b for b in bugs}
+        with _LOCK:
+            seen = set(_STATE["claims"].keys())
+            retry_ids = [k for k, c in _STATE["claims"].items()
+                         if c.get("state") == "need_manual" and k in by_id
+                         and str(by_id[k].get("status") or "") == "active"]
+        for bug in bugs:
+            bid = str(bug.get("id") or "")
+            if not bid or bid in seen or not _claimable(bug, profile):
+                continue
+            _route_one(bug, profile, cfg)
+            claimed += 1
+        # need_manual 重排查（bug 仍激活才出现在列表里）
+        for bid in retry_ids:
+            with _LOCK:
+                if _STATE["claims"].get(bid, {}).get("state") != "need_manual":
+                    continue
+            _route_one(by_id[bid], profile, cfg, notify=False)
+    return claimed
+
+
+def _poll(force=False):
+    """执行一次扫描；同一进程内只允许一个扫描实例。"""
+    if not _SCAN_LOCK.acquire(blocking=False):
+        return {"ok": False, "claimed": 0, "reconciled": False,
+                "error": "扫描正在进行，请稍候", "skipped": "in_progress"}
+    try:
+        return _poll_unlocked(force=force)
+    finally:
+        _SCAN_LOCK.release()
+
+
+def _poll_unlocked(force=False):
+    """一次完整轮询：对账回写 + （到点/强制时）扫描认领。异常不外抛。"""
+    _ensure_loaded()
+    with _LOCK:
+        cfg = _cfg()
+    out = {"ok": True, "claimed": 0, "reconciled": True, "error": ""}
+    try:
+        _reconcile(cfg)
+    except Exception:
+        log.exception("zentao: 对账异常")
+    if not force:
+        if not cfg.get("poll_enabled"):
+            return {"ok": True, "claimed": 0, "reconciled": True, "error": "",
+                    "skipped": "poll_disabled"}
+        with _LOCK:
+            nxt = _STATE.get("next_scan") or ""
+        if nxt:
+            try:
+                if datetime.fromisoformat(nxt) > datetime.now():
+                    return {"ok": True, "claimed": 0, "reconciled": True,
+                            "error": "", "skipped": "not_due"}
+            except ValueError:
+                pass
+    err = ""
+    try:
+        out["claimed"] = _scan(cfg)
+    except ZenError as e:
+        err = str(e)
+    except Exception as e:
+        err = "扫描失败：%s" % e
+        log.warning("zentao: %s", err)
+    now = datetime.now()
+    delay = timedelta(minutes=max(INTERVAL_MIN, int(cfg.get("interval_minutes") or INTERVAL_DEFAULT)))
+    if err:
+        delay = timedelta(minutes=RETRY_DELAY_MIN)
+    with _LOCK:
+        _STATE["last_scan"] = now.strftime("%Y-%m-%d %H:%M:%S")
+        _STATE["next_scan"] = (now + delay).strftime("%Y-%m-%d %H:%M:%S")
+        _STATE["last_error"] = err
+        _save_locked()
+    out["error"] = err
+    out["ok"] = not err
+    return out
+
+
+def fire_due():
+    """automation._tick 每拍调用：内部自节流，没到点/未启用时零开销返回。"""
+    _ensure_loaded()
+    with _LOCK:
+        if not (_STATE.get("config") or {}).get("poll_enabled"):
+            return None
+    return _poll(force=False)
+
+
+def scan_now():
+    """手动「立即扫描」：绕过轮询闸与节流。"""
+    return _poll(force=True)
+
+
+def _boot_reconcile():
+    """服务重启后补一次对账：上个进程死亡窗口里的终态变化别等下个 tick
+    （#27697 案：服务死窗 7 小时，失败转派被拖到重启后才发现）。"""
+    try:
+        _ensure_loaded()
+        with _LOCK:
+            cfg = _cfg()
+        _reconcile(cfg)
+    except Exception:
+        log.exception("zentao: 启动补对账失败")
+
+
+def start():
+    """服务启动接线：加载状态（含老配置迁移）。"""
+    global _BOOT_TIMER
+    n = load()
+    with _LOCK:
+        cfg = _STATE.get("config") or {}
+        if cfg.get("poll_enabled") and not _STATE.get("next_scan"):
+            _STATE["next_scan"] = time.strftime("%Y-%m-%d %H:%M:%S")
+            _save_locked()
+        pending = any(c.get("state") == "fixing" for c in _STATE["claims"].values())
+    if pending:
+        if _BOOT_TIMER is not None:
+            _BOOT_TIMER.cancel()
+        _BOOT_TIMER = threading.Timer(3.0, _boot_reconcile)
+        _BOOT_TIMER.daemon = True
+        _BOOT_TIMER.start()
+    return n
+
+
+# ---------------------------------------------------------------- 配置管理
+
+def _norm_profile(p, idx):
+    if not isinstance(p, dict):
+        raise ValueError("产品档案 #%d 必须是对象" % idx)
+    try:
+        pid = int(p.get("product") or 0)
+    except (TypeError, ValueError):
+        raise ValueError("产品档案 #%d 的产品 ID 必须是整数" % idx)
+    if pid <= 0:
+        raise ValueError("产品档案 #%d 缺产品 ID" % idx)
+    prof = dict(_PROFILE_DEFAULTS)
+    prof["product"] = pid
+    prof["assigned_to"] = str(p.get("assigned_to") or "").strip()
+    try:
+        prof["severity_cap"] = max(0, min(4, int(p.get("severity_cap") or 0)))
+    except (TypeError, ValueError):
+        prof["severity_cap"] = 0
+    our = [s for s in SIDES if s in (p.get("our_sides") or [])]
+    prof["our_sides"] = our
+    repos = {"backend": dict(_REPO_DEFAULTS), "frontend": dict(_REPO_DEFAULTS)}
+    raw_repos = p.get("repos") if isinstance(p.get("repos"), dict) else {}
+    for side in SIDES:
+        r = raw_repos.get(side) if isinstance(raw_repos.get(side), dict) else {}
+        wd = str(r.get("workdir") or "").strip()
+        if wd:
+            rp = Path(wd).expanduser()
+            if not rp.is_absolute():
+                raise ValueError("产品 %d 的%s工作目录必须是绝对路径" % (pid, SIDE_CN[side]))
+            wd = str(rp)
+        repos[side] = {"workdir": wd,
+                       "git_rev": str(r.get("git_rev") or "").strip(),
+                       "verify_command": str(r.get("verify_command") or "").strip()}
+    # 工作目录留空合法：扫描时回落「默认保存路径」（_launch_fix 同口径），不在此拦截
+    prof["repos"] = repos
+    hints = {}
+    raw_hints = p.get("repo_hints") if isinstance(p.get("repo_hints"), dict) else {}
+    for side in SIDES:
+        hints[side] = str(raw_hints.get(side) or "").strip()
+    prof["repo_hints"] = hints
+    owners = {}
+    raw_owners = p.get("owners") if isinstance(p.get("owners"), dict) else {}
+    for k in ("backend", "frontend", "not_ours"):
+        owners[k] = str(raw_owners.get(k) or "").strip()
+    prof["owners"] = owners
+    routes = []
+    for r in (p.get("module_routes") or []):
+        if not isinstance(r, dict):
+            continue
+        try:
+            mid = int(r.get("module") or 0)
+        except (TypeError, ValueError):
+            raise ValueError("产品 %d 的模块路由：模块 ID 必须是整数" % pid)
+        if mid <= 0:
+            continue
+        side = str(r.get("side") or "")
+        if side not in TRIAGE_SIDES:
+            raise ValueError("产品 %d 的模块路由 side 必须是 %s 之一"
+                             % (pid, "/".join(TRIAGE_SIDES)))
+        routes.append({"module": mid, "side": side,
+                       "account": str(r.get("account") or "").strip()})
+    prof["module_routes"] = routes
+    return prof
+
+
+def _norm_profiles(v):
+    if v is None:
+        return None
+    if not isinstance(v, list):
+        raise ValueError("product_profiles 必须是数组")
+    out, seen = [], set()
+    for i, p in enumerate(v):
+        prof = _norm_profile(p, i)
+        if prof["product"] in seen:
+            raise ValueError("产品 %d 配置重复" % prof["product"])
+        seen.add(prof["product"])
+        out.append(prof)
+    return out
+
+
+def save_config(patch):
+    """部分更新配置（password 缺省或空串=不改；product_profiles 整体替换）。
+    校验失败抛 ValueError。返回脱敏视图的 config。"""
+    _ensure_loaded()
+    patch = patch if isinstance(patch, dict) else {}
+    with _LOCK:
+        cfg = _cfg()
+        for k in _UPDATABLE:
+            if k not in patch:
+                continue
+            v = patch[k]
+            if k == "base_url":
+                v = str(v or "").strip().rstrip("/")
+                if v and not v.startswith(("http://", "https://")):
+                    raise ValueError("禅道地址必须以 http:// 或 https:// 开头")
+            elif k == "account":
+                v = str(v or "").strip()
+            elif k == "password":
+                v = str(v or "")
+                if not v:
+                    continue
+            elif k == "product_profiles":
+                v = _norm_profiles(v)
+                if v is None:
+                    continue
+            elif k == "interval_minutes":
+                try:
+                    v = max(INTERVAL_MIN, min(INTERVAL_MAX, int(v)))
+                except (TypeError, ValueError):
+                    raise ValueError("interval_minutes 必须是 %d-%d 的整数"
+                                     % (INTERVAL_MIN, INTERVAL_MAX))
+            elif k in ("auto_resolve", "auto_merge", "triage_ai", "poll_enabled"):
+                v = bool(v)
+            cfg[k] = v
+        _STATE["config"] = cfg
+        if cfg.get("poll_enabled"):
+            _STATE["next_scan"] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+        _save_locked()
+    return view()["config"]
+
+
+def bug_web_base(cfg=None):
+    """给前端拼 bug 详情页用的 web 根：探测出的有效地基（含自动补的 /zentao 子路径）
+    优先，没探测过回落用户填的原始地址。返回 stripped 根或空串。"""
+    c = dict(cfg) if cfg else _cfg()
+    hit = resolved_base_url(c.get("base_url"))
+    if hit:
+        return hit
+    b = re.sub(r"/api\.php/v1$", "", str(c.get("base_url") or "").strip().rstrip("/"))
+    return b if b.startswith(("http://", "https://")) else ""
+
+
+def bug_link_style(cfg=None):
+    """bug 详情链接形态：探测缓存优先，没探出来按伪静态（官方默认形态）。"""
+    c = dict(cfg) if cfg else _cfg()
+    b = str(c.get("base_url") or "").strip().rstrip("/")
+    return _WEBSTYLE.get(b) or "pathinfo"
+
+
+def view():
+    """前端视图：配置脱敏（password 只回是否已设）+ claims 列表（新在前）。"""
+    _ensure_loaded()
+    with _LOCK:
+        cfg = dict(_cfg())
+        has_pw = bool(cfg.get("password"))
+        cfg["password"] = ""
+        cfg["has_password"] = has_pw
+        bug_base = bug_web_base(cfg)
+        claims = sorted(_STATE["claims"].values(),
+                        key=lambda c: str(c.get("claimed_at") or ""), reverse=True)
+        return {"config": cfg,
+                "bug_base": bug_base,
+                "bug_style": bug_link_style(cfg),
+                "claims": [dict(c) for c in claims],
+                "last_scan": _STATE.get("last_scan") or "",
+                "next_scan": _STATE.get("next_scan") or "",
+                "last_error": _STATE.get("last_error") or ""}
+
+
+def _test_reset():
+    """测试钩子：清空内存状态（配合重绑 _FILE + TUTTI_DATA 临时目录用）。"""
+    with _LOCK:
+        _STATE["config"] = dict(_CFG_DEFAULTS)
+        _STATE["claims"] = {}
+        _STATE["last_scan"] = _STATE["next_scan"] = _STATE["last_error"] = ""
+        _reset_token()
+        _APIBASE.clear()
+        _WEBSTYLE.clear()
+        _MODE.clear()
+        _OLD.update(api="", sid="", at=0.0)
+        _OLD_FORM["form"] = ""
+        _REST_LAST_ERR["msg"] = ""
+        _IMG_CACHE.clear()
+        globals()["_LOADED"] = True