crawlrs 0.2.0

crawlrs - Enterprise-grade web scraping and crawling platform built with Rust
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704

# 📘 Crawlrs API Reference

<div align="center">

![API Version](https://img.shields.io/github/v/release/Kirky-X/crawlrs?label=api&style=for-the-badge)
![Base URL](https://img.shields.io/badge/base%20URL-http://localhost:8899-green?style=for-the-badge)
![License](https://img.shields.io/github/license/Kirky-X/crawlrs?label=license&style=for-the-badge)

**Version:** 0.2.0 | **Base URL:** `http://localhost:8899` | **Updated:** 2026-11-15

</div>

---

## 📋 目录

- [🔐 Authentication](#-authentication)
- [📄 Common Response Format](#-common-response-format)
- [⚠️ Errors](#️-errors)
- [🌐 Public Endpoints](#-public-endpoints)
  - [Health Check](#health-check-liveness)
  - [Readiness Check](#readiness-check)
  - [Garrison Health Probes](#garrison-health-probes-auth-builds-only)
  - [OpenAPI Specification](#openapi-specification)
  - [Get Version](#get-version)
  - [Get Metrics](#get-metrics)
- [🔒 Protected Endpoints](#-protected-endpoints)
  - [Scrape API](#scrape-api)
  - [Crawl API](#crawl-api)
  - [Search API](#search-api)
  - [Extract API](#extract-api)
  - [Task API](#task-api)
  - [Team API](#team-api)
  - [Webhook API](#webhook-api)
  - [Audit API](#audit-api)
  - [Admin API](#admin-api)
- [⏱️ Rate Limiting](#️-rate-limiting)
- [🪝 Webhooks](#-webhooks)
- [🧩 SDK API](#-sdk-api)
- [💻 SDK Examples](#-sdk-examples)
- [✅ Best Practices](#-best-practices)
- [📋 Changelog](#-changelog)
- [📞 Support](#-support)

---

## 🔐 Authentication

All protected endpoints require authentication using an API key in the `Authorization` header:

```http
Authorization: Bearer YOUR_API_KEY
```

> **0.2.0 起(`garrison-auth-migration`):** Bearer token 由 **garrison v0.9** 校验,格式为 `garrison_key_id.garrison_secret`(明文 key 仅在签发时返回一次)。旧的 SHA-256 `api_key_hash` 已作废,需经 garrison 重新领取。

### Garrison 认证

**认证流程:**

1. 客户端在 `Authorization: Bearer` 头中传入 garrison 签发的 key(格式 `garrison_key_id.garrison_secret`)
2. `auth_middleware_inner` 调用 `GarrisonUtil::check_api_key` 校验 key、过期、状态、RBAC 权限
3. `auth_bridge::bridge_to_auth_state` 将 garrison permissions 通过 `map_perms_to_scope` 映射为 `ApiKeyScope`,注入 `AuthState`
4. 401(无效 key)/ 429(暴力破解锁定)由 garrison `firewall-bruteforce` 触发(默认 5 次失败/60 秒窗口/300 秒锁定)

**RBAC 权限映射(`tenant_id=0`,所有 team 共享,admin 蕴含 read+write):**

| garrison 权限 | 映射到 `ApiKeyScope` |
|---------------|----------------------|
| `crawlrs:admin` | `read=true, write=true, admin=true` |
| `crawlrs:write` | `write=true` |
| `crawlrs:read` | `read=true` |

全部通过 `ApiKeyScope::with_custom_limits` 构造(`DEFAULT_SEARCH_LIMIT=100`, `DEFAULT_SCRAPE_LIMIT=50`),不调用 `full_access()`——避免无限制 `u32::MAX` 与项目配额语义冲突。

**JWT 配置:**

| 环境变量 | 描述 | 必填 |
|----------|------|------|
| `CRAWLRS__AUTH__JWT_SECRET` | HS256 JWT 密钥,≥32 字节;弱密钥拒绝启动 | 是(`auth` feature 开启时) |

### API Key Scopes

API keys carry an `ApiKeyScope` with three boolean permissions plus two hourly limits (`src/domain/auth/mod.rs`):

| Field | Description |
|--------|-------------|
| `read` | Permission to access read-only endpoints |
| `write` | Permission to access write endpoints |
| `admin` | Permission to access administrative endpoints(蕴含 read + write) |
| `search_limit` | Maximum search requests per hour |
| `scrape_limit` | Maximum scrape requests per hour |

> **重要现状(0.2.0-rc.2):** `scope_middleware` 虽已实现(`src/presentation/middleware/auth_middleware.rs`),但**未挂载到生产路由**——仅存在于测试 Router 装配。除 `POST /v1/admin/api-keys` 在 handler 内部显式校验 `Admin` 权限外,其余端点目前不校验 API Key scope,仅校验 key 有效性。请勿误以为 RBAC 已在端点层生效。

### Auth Disabled Behavior (auth-off)

> **Conditional:** When the `auth` feature is disabled (via `--no-default-features` or explicit feature selection), the real `auth_middleware_inner` is not compiled. `default_identity_middleware` injects a prebuilt `AuthState` template instead — but **默认拒绝**:未显式 opt-in 时,所有 protected 请求一律返回 `401`。

| Behavior | When `auth` Enabled (default) | When `auth` Disabled |
|----------|------------------------------|---------------------|
| `Authorization` header | Required (401 if missing/invalid; 429 on brute-force lockout) | **Ignored, but requests still 401** until `allow_unauthenticated_protected()` is called |
| Opt-in | N/A | Must explicitly call `auth_middleware::allow_unauthenticated_protected()` (one-time `error!` log declares the security downgrade) |
| `AuthState.team_id` | Resolved via garrison → `api_key_id→team_id` 反查 + `TEAM_ID_CACHE` LRU | Fixed to `DEFAULT_TEAM_ID` (`domain::id::Id::from_u128(1)`) |
| `AuthState.api_key_id` | Resolved from garrison key | Fixed to `DEFAULT_API_KEY_ID` (`domain::id::Id::from_u128(2)`) |
| `AuthState.scope` | `map_perms_to_scope(garrison permissions)` | `ApiKeyScope::denied()` (read/write/admin all `false`, limits 0) — 受限匿名身份,非 `full_access()` |
| Token cache | garrison 自管 oxcache;crawlrs 侧 `TEAM_ID_CACHE` 缓存映射 | N/A — fixed identity, no caching |
| Rate limit lockout | garrison `firewall-bruteforce` (5/60s/300s) | Not tracked (but `rate-limit` feature may still apply per-IP limits) |

> 源码依据:`src/bootstrap/routes.rs::build_default_identity_template`(模板携带 `ApiKeyScope::denied()`)与 `src/presentation/middleware/auth_middleware.rs::default_identity_middleware`(默认 401,opt-in 后注入模板)。opt-in 放行后的身份仍是 denied 受限身份,业务层若按权限判断将拒绝服务。

**Build commands:**
```bash
# Auth disabled (single-tenant, no auth)
cargo build --release --no-default-features

# Auth enabled (default, includes garrison)
cargo build --release --features default
```

### 迁移指南(0.2.0 garrison-auth-migration)

**对现有 API 客户端的影响:**

- 旧 API Key(基于 SHA-256 `api_key_hash`)全部作废,必须经 garrison 重新领取
- 客户端无需改动 `Authorization: Bearer` 调用方式,仅需替换为新的 `garrison_key_id.garrison_secret`
- 401/429 响应语义不变,但触发逻辑由 garrison `firewall-bruteforce` 接管

**迁移已完成:**

- 旧 API Key 已作废,客户端须使用 `Authorization: Bearer <garrison_key_id>.<garrison_secret>` 格式
- 新 team 通过 `POST /v1/admin/api-keys` 签发 API Key,无需运维工具介入

详细架构说明参见 [ARCHITECTURE.md → Garrison 认证引擎](ARCHITECTURE.md#garrison-认证引擎)。

---

## 📄 Common Response Format

All API responses follow this unified structure:

### Success Response

```json
{
  "success": true,
  "data": {
    // Response data here
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

### Success Response with Pagination

```json
{
  "success": true,
  "data": {
    // Response data here
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> **当前无端点使用 `meta` 分页包封。** `ApiResponse` 结构体与 `success_response_with_meta` 构造器在 `src/presentation/handlers/response_builder.rs` 有定义(携带 `PaginationMeta`),但生产 handler 无任何调用点;列表端点(如 `GET /v1/webhooks`、audit logs)的分页信息直接平铺在 `data` 内。上述 `meta` 形态仅为保留的库能力,不是现行 API 契约。

### Error Response

```json
{
  "success": false,
  "error": {
    "code": "VALIDATION_ERROR",
    "message": "Detailed error message"
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

### Rate Limit Error Response

限流拒绝(429)由各 handler 内 `check_rate_limit` 经标准 `ApiResponse` 包封返回,`error` 为 `{code,message}` 对象:

```json
{
  "success": false,
  "error": {
    "code": "RATE_LIMITED",
    "message": "Rate limit exceeded: <reason>"
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

当命中 `RetryAfter` 结果时额外带 `retry_after_seconds` 字段:

```json
{
  "success": false,
  "error": {
    "code": "RATE_LIMITED",
    "message": "Rate limit exceeded, please retry after 60 seconds"
  },
  "retry_after_seconds": 60,
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> 源码依据:`src/presentation/helpers/rate_limit_helper.rs::check_rate_limit`(`Denied` → `errors::too_many_requests`,`RetryAfter` → `errors::too_many_requests_with_retry`,后端错误 fail-closed 503 → `errors::service_unavailable`)。响应头仅设 `Retry-After`,无 `X-RateLimit-*` 系列头。

### Response Fields

| Field | Type | Description |
|-------|------|-------------|
| `success` | boolean | Whether the request was successful |
| `data` | object | Response data (only present on success) |
| `error` | object | Error details (only present on failure) |
| `error.code` | string | Error code for programmatic handling |
| `error.message` | string | Human-readable error message |
| `timestamp` | string | Response timestamp in RFC3339 format |

---

## ⚠️ Errors

### HTTP Status Codes

| Code | Description |
|------|-------------|
| 200 | Success |
| 201 | Created |
| 400 | Bad Request - Invalid parameters |
| 401 | Unauthorized - Missing or invalid API key |
| 403 | Forbidden - Insufficient permissions |
| 429 | Too Many Requests - Rate limit exceeded |
| 422 | Unprocessable Entity - Validation error |
| 500 | Internal Server Error |

### Error Codes

生产错误码清单(`src/common/error.rs::CrawlRsError::error_code` 与 `src/presentation/handlers/response_builder.rs::error_codes`):

| Error Code | HTTP Status | Description |
|------------|-------------|-------------|
| `VALIDATION_ERROR` | 400 | Invalid request parameters |
| `NOT_FOUND` | 404 | Resource not found |
| `UNAUTHORIZED` | 401 | Missing or invalid API key |
| `AUTHENTICATION_ERROR` | 401 | Authentication failed (e.g. webhook signature verification) |
| `FORBIDDEN` | 403 | Insufficient permissions |
| `RATE_LIMITED` | 429 | Rate limit exceeded |
| `UNPROCESSABLE_ENTITY` | 422 | Structural validation error (field missing / wrong type / unknown field) |
| `INTERNAL_ERROR` | 500 | Internal server error |
| `SERVICE_UNAVAILABLE` | 503 | Service unavailable |
| `TASK_ERROR` | 500 | Task processing error |
| `ENGINE_ERROR` | 502 | Scraping engine error |
| `JSON_ERROR` | 400 | JSON serialization/deserialization error |
| `IO_ERROR` | 500 | Internal I/O error |
| `CONFIGURATION_ERROR` | 500 | Configuration error |
| `QUOTA_EXCEEDED` | 402 | Quota exceeded |
| `DATABASE_ERROR` | 500 | Database error(经 `CrawlRsError` 路径间接产出) |
| `CACHE_ERROR` | 503 | Cache error(经 `CrawlRsError` 路径间接产出) |
| `EXTERNAL_SERVICE_ERROR` | 502 | External service error(经 `CrawlRsError::Network` 间接产出) |
| `TIMEOUT` | 504 | Request timeout(经 `CrawlRsError::Timeout` 间接产出) |

> `CONFLICT` / `PRECONDITION_FAILED` / `FEATURE_DISABLED` 在 `error_codes` 模块有常量定义,但无生产发射点;feature 未编译时对应路由根本不注册,请求得到的是普通 404(`NOT_FOUND`)。

> `/metrics`:公开 Prometheus 文本端点(免认证、免限流),仅暴露运行态指标,
> 不含租户 ID/webhook URL/凭证等敏感标签。生产建议在反向代理层做网络 allowlist;
> 指标传输已改为进程内 recorder(不再监听 9100)。


---

## 🌐 Public Endpoints

### Health Check (Liveness)

Check if the API process is running. Always returns 200 when the process is alive — dependencies are not probed here.

**Endpoint:** `GET /health`

**Response:**
```json
{
  "status": "healthy",
  "version": "0.2.0"
}
```

### Readiness Check

Kubernetes readiness probe: probes PostgreSQL (`SELECT 1`, 3s timeout) and the cache service (`get("__readiness_probe__")`, 1s timeout). Returns 200 with `status: "ready"` when all dependencies are up, 503 with `status: "not_ready"` otherwise.

**Endpoint:** `GET /ready`

**Response (200):**
```json
{
  "status": "ready",
  "checks": {
    "database": { "status": "up" },
    "cache": { "status": "up" },
    "database_pool": { "status": "ok" }
  },
  "version": "0.2.0"
}
```

**Response (503):**
```json
{
  "status": "not_ready",
  "checks": {
    "database": { "status": "down", "error": "..." },
    "cache": { "status": "up" },
    "database_pool": { "status": "ok" }
  },
  "version": "0.2.0"
}
```

> `checks` 含 `database`、`cache` 与 `database_pool`(连接池结构化快照:容量/空闲/等待,仅诊断用途,不影响就绪判定)。各探测失败标记 `status: "down"` 并附 `error`(超时为 `"timeout"`)。

### Garrison Health Probes (`auth` builds only)

When the `auth` feature is enabled, garrison's `health_routes` are merged into the public router (`src/bootstrap/routes.rs::create_public_routes`), exposing two additional paths:

- `GET /health/live` — liveness probe, always 200
- `GET /health/ready` — readiness probe aggregating the garrison `HealthRegistry` checks; `overall` Unhealthy returns 503 (K8s traffic-removal semantics carried by the library)

These paths do not collide with `/health` and `/ready`. `auth`-off builds do not compile these routes.

### OpenAPI Specification

**Endpoint:** `GET /openapi.json` (requires the `openapi` feature; registered at the root path by `src/presentation/sdk/mod.rs`)

**Response:** OpenAPI 3.1 document (JSON), generated by sdforge from all registered SDK endpoints.

### Get Version

Get the current API version.

**Endpoint:** `GET /v1/version`

**Response:**
```text
0.2.0
```

### Get Metrics

**Endpoint:** `GET /metrics`

> **⚠️ 当前为占位实现。** `metrics_handler` 返回 JSON 包封,其中 `metrics` 字段是一段**硬编码的 Prometheus 风格文本**——`version` 固定为 `1.0.0`、uptime 固定为 `3600`,并非运行时真实指标,也不是 Prometheus 文本格式聚合端点。权威的真实指标清单(采集点与标签定义)见 [PERFORMANCE.md](PERFORMANCE.md)。`/metrics` 的权威抓取格式(可直接被 Prometheus scraper 消费)尚未实现。

**Response:**
```json
{
  "status": "ok",
  "metrics": "# HELP app_info Information about the application\n# TYPE app_info gauge\napp_info{version=\"1.0.0\"} 1\n# HELP app_uptime_seconds Application uptime in seconds\n# TYPE app_uptime_seconds counter\napp_uptime_seconds 3600"
}
```

---

## 🔒 Protected Endpoints

### Scrape API

Scrape a single web page.

#### Create Scrape Task

**Endpoint:** `POST /v1/scrape`

**Request Body:**
```json
{
  "url": "https://example.com",
  "formats": ["markdown", "html"],
  "include_tags": ["h1", "h2", "p"],
  "exclude_tags": ["script", "style"],
  "webhook": "https://your-webhook.com/callback",
  "extraction_rules": {
    "title": {
      "selector": "h1",
      "attr": "text"
    }
  },
  "actions": [
    {
      "type": "wait",
      "milliseconds": 1000
    },
    {
      "type": "click",
      "selector": ".load-more"
    }
  ],
  "options": {
    "headers": {
      "User-Agent": "Mozilla/5.0..."
    },
    "wait_for": 2000,
    "timeout": 30,
    "js_rendering": false,
    "screenshot": true,
    "screenshot_options": {
      "full_page": true,
      "selector": "#main",
      "quality": 90,
      "format": "png"
    },
    "mobile": false,
    "proxy": "http://proxy.example.com:8080",
    "skip_tls_verification": false,
    "needs_tls_fingerprint": false,
    "use_fire_engine": false
  },
  "metadata": {
    "custom_key": "custom_value"
  },
  "sync_wait_ms": 5000
}
```

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `url` | string | Yes | Target URL (http/https only) |
| `formats` | array | No | Output formats: `markdown`, `html`, `text` |
| `include_tags` | array | No | HTML tags to include in output |
| `exclude_tags` | array | No | HTML tags to exclude from output |
| `webhook` | string | No | Webhook URL for completion notification |
| `extraction_rules` | object | No | CSS selector extraction rules |
| `extraction_prompt` | string | No | LLM prompt for data extraction (free text) |
| `extraction_schema` | object | No | JSON Schema for structured LLM extraction |
| `actions` | array | No | Page interaction actions |
| `options` | object | No | Scraping options (see [Scrape API](#scrape-api)) |
| `metadata` | object | No | Custom metadata for the task |
| `sync_wait_ms` | integer | No | Wait time for synchronous response (max 30000) |

<a id="scrape-options-parameters"></a>

**Options Parameters:**

| Option | Type | Default | Description |
|--------|------|---------|-------------|
| `headers` | object | empty | Custom HTTP headers (e.g. `User-Agent`) |
| `wait_for` | integer | — | Wait hint in milliseconds (DTO validation max 60000, but the use case clamps `ms.min(10_000)` — values above 10000 are silently truncated to 10000). Mapped to the `DomStable` wait strategy (DOM unchanged for this duration ⇒ ready); `None` uses the engine default (`WaitFor::NetworkIdle`, CDP network-event based). Mutually exclusive with `wait_for_response_url` / `wait_for_response_body`. See [ARCHITECTURE.md → WaitFor](ARCHITECTURE.md#waitfor-策略). |
| `wait_for_response_url` | string | — | Response-level wait: resolve when a response URL matches this regex ([regex crate](https://docs.rs/regex) syntax, e.g. `api/v\d+/items`). Validated at request time — an invalid regex is rejected before the task is enqueued (`wait_for`-style silent truncation does not apply). Engine support: Playwright (CDP response events); moli returns `Unsupported` (its CDP Network event surface is unavailable — see `scripts/poc/moli/network-probe-output.txt`). On timeout the request fails with the standard wait-timeout error. |
| `wait_for_response_body` | string | — | Response-level wait: resolve when a response body contains this substring (e.g. `"status":"ready"`). Per-response matching is capped at 10 MiB (bodies beyond the cap are skipped, not buffered). Same exclusivity, engine support and timeout semantics as `wait_for_response_url`. |
| `timeout` | integer | 30 | Request timeout in seconds |
| `js_rendering` | boolean | false | Enable JavaScript rendering |
| `screenshot` | boolean | false | Capture screenshot |
| `screenshot_options` | object | — | Screenshot config (`full_page`, `selector`, `quality`, `format`) |
| `mobile` | boolean | false | Emulate mobile device |
| `proxy` | string | — | Proxy URL (e.g. `http://proxy:8080`) |
| `skip_tls_verification` | boolean | false | Skip TLS certificate verification |
| `needs_tls_fingerprint` | boolean | false | Require TLS fingerprint adversarial engine (`engine-tls-fingerprint` feature) |
| `use_fire_engine` | boolean | false | Force use FlareSolverr engine |
| `cache_mode` | string | `"enabled"` | Cache read/write mode: `enabled` (default, normal R/W), `disabled` (cache fully off), `read_only` (hit returns, miss fetches without writing), `bypass` (skip read, normal write — emergency untrusted-cache scenario) |
| `bypass_cache` | boolean | false | Emergency cache bypass shortcut. `true` overrides `cache_mode` to `bypass` (skip cache read, normal write-back). Use when runtime cache data is untrusted |
| `only_main_content` | boolean | — | Extract main content before Markdown conversion (strips nav/footer/sidebar/ads). Requires the `markdown` format |
| `respect_robots` | boolean | — | Honor the target site's robots.txt for this request (falls back to global `robots.scrape_respect_robots`, default off) |
| `conditional_request` | boolean | — | Incremental fetch: send `If-None-Match`/`If-Modified-Since` when a prior result exists; a `304` reuses the historical content (`x-crawlrs-not-modified: true`) |
| `block_ads` | boolean | false | Block common ad-domain requests in browser engines |
| `block_media` | boolean | false | Block image/video/audio media loading in browser engines (speeds up text-only scraping) |
| `needs_mllm` | boolean | false | Require MLLM visual analysis of the screenshot. Requires the `engine-mllm` feature: builds without it reject the request **at submission time** with an explicit validation error (`needs_mllm requires the \`engine-mllm\` feature`) instead of silently degrading to a non-MLLM engine; the worker repeats the check for paths that bypass the `create_scrape` use case |

> `session_id` remains an internal engine field (browser-session reuse is not exposed per-request yet). Unknown fields are still rejected with 422 via `deny_unknown_fields`.

**Action Types:**

| Type | Parameters | Description |
|-------|-----------|-------------|
| `wait` | `milliseconds` | Wait for specified time |
| `click` | `selector` | Click element matching selector |
| `scroll` | `direction` | Scroll page (up/down) |
| `screenshot` | `full_page` | Take screenshot |
| `input` | `selector`, `text` | Input text into element |

**Response (Success):**
```json
{
  "success": true,
  "data": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "url": "https://example.com",
    "credits_used": 1
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> `credits_used` 固定为 1(每次抓取任务入队时扣减 1 credit;入队失败自动退还)。状态码:入队成功为 `201 Created`;`sync_wait_ms > 0` 且等待超时时为 `202 Accepted`(body 结构不变,仍含任务 `id`)。

#### Get Scrape Status

**Endpoint:** `GET /v1/scrape/{id}`

**Parameters:**
- `id` (path) - Task UUID

**Response:**
```json
{
  "success": true,
  "data": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "status": "completed",
    "url": "https://example.com",
    "created_at": "2025-01-15T12:00:00",
    "completed_at": "2025-01-15T12:00:03",
    "result": {
      "content": "<html>...</html>",
      "status_code": 200,
      "content_type": "text/html",
      "response_time_ms": 1200,
      "headers": { "content-type": "text/html" },
      "meta_data": {},
      "screenshot": null,
      "created_at": "2025-01-15T12:00:03"
    },
    "metadata": null,
    "error": null
  },
  "timestamp": "2025-01-15T12:00:03+00:00"
}
```

> 响应为平铺的 `ScrapeStatusResponseDto`(无 `task` 外层键)。`result` 仅在任务 `completed` 时出现;敏感响应头(`set-cookie` / `authorization` 等)在返回前脱敏为 `[REDACTED]`。`error` 仅在任务 `failed` 时出现。
>
> 可选字段 `js_degraded`(boolean):仅当请求要求 JS 渲染但内存水位(`engines.memory_scheduler`,`engine-moli` 特性)触发静态回退时以 `true` 出现——正常响应与旧缓存条目均不含该字段。

#### Cancel Scrape

Cancel a single scrape task by ID. Only the owning team can cancel; the task is marked `cancelled`.

**Endpoint:** `POST /v1/scrape/{id}/_cancel`

**Parameters:**
- `id` (path) - Task UUID

**Response:**
```json
{
  "success": true,
  "data": {
    "message": "Scrape task cancelled"
  },
  "timestamp": "2025-01-15T12:00:05+00:00"
}
```

---

### Crawl API

Crawl multiple pages from a starting URL.

#### Create Crawl Task

**Endpoint:** `POST /v1/crawl`

**Request Body:**
```json
{
  "url": "https://example.com",
  "name": "my-crawl",
  "config": {
    "max_depth": 2,
    "max_pages": 100,
    "include_patterns": ["/blog/.*"],
    "exclude_patterns": ["/admin/.*"],
    "strategy": "bfs",
    "crawl_delay_ms": 1000,
    "max_concurrency": 4,
    "proxy": "http://proxy.example.com:8080"
  },
  "sync_wait_ms": 10000
}
```

> 请求体经 `#[serde(deny_unknown_fields)]` 校验:未列出的字段(如 `formats`、`webhook`、`options`、`follow_links`)会被 422 拒绝。

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `url` | string | Yes | Starting URL (http/https only) |
| `name` | string | No | Task name (1-255 chars) |
| `config` | object | Yes | Crawl configuration (see below) |
| `sync_wait_ms` | integer | No | Wait time for synchronous response (0-30000, default 5000) |
| `expires_at` | string | No | Task expiry time (RFC3339) |
| `sitemap_seed` | boolean | No | Use sitemap as URL discovery source (requires `CRAWLRS__SITEMAP__ENABLED`) |
| `sitemap_url` | string | No | Explicit sitemap URL (defaults to `<origin>/sitemap.xml`) |

**`config` object (`CrawlConfigDto`):**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `max_depth` | integer | Yes | Maximum crawl depth (required; DTO max 100, but the handler tightens this to ≤ 5 — `max_depth > 5` returns 422) |
| `include_patterns` | array | No | Regex patterns for URLs to include |
| `exclude_patterns` | array | No | Regex patterns for URLs to exclude |
| `strategy` | string | No | Traversal strategy: `bfs` (default) or `dfs` |
| `crawl_delay_ms` | integer | No | Per-host crawl delay (max 86400000) |
| `max_concurrency` | integer | No | Max concurrent pages (max 50) |
| `max_pages` | integer | No | Maximum pages to crawl (default 1000, max 100000; exceeded stops early) |
| `proxy` | string | No | Proxy URL |
| `headers` | object | No | Custom HTTP headers |
| `extraction_rules` | object | No | CSS selector extraction rules |
| `extraction_prompt` | string | No | LLM prompt for data extraction on each crawled page |
| `extraction_schema` | object | No | JSON Schema for structured LLM extraction |

**Response (Success):**
```json
{
  "success": true,
  "data": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "team_id": "660e8400-e29b-41d4-a716-446655440001",
    "name": "my-crawl",
    "root_url": "https://example.com",
    "url": "https://example.com",
    "status": "queued",
    "config": { "max_depth": 2 },
    "total_tasks": 0,
    "completed_tasks": 0,
    "failed_tasks": 0,
    "created_at": "2025-01-15T12:00:00Z",
    "updated_at": "2025-01-15T12:00:00Z",
    "completed_at": null
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> 响应为裸 `Crawl` 域模型(无 `credits_used` 字段,无 `task` 包装)。状态码:创建成功为 `201 Created`;`sync_wait_ms > 0` 且等待超时时为 `202 Accepted`。

#### Get Crawl Status

**Endpoint:** `GET /v1/crawl/{id}`

**Parameters:**
- `id` (path) - Crawl UUID

**Response:**
```json
{
  "success": true,
  "data": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "team_id": "660e8400-e29b-41d4-a716-446655440001",
    "name": "my-crawl",
    "root_url": "https://example.com",
    "url": "https://example.com",
    "status": "processing",
    "config": { "max_depth": 2 },
    "total_tasks": 100,
    "completed_tasks": 45,
    "failed_tasks": 2,
    "created_at": "2025-01-15T12:00:00Z",
    "updated_at": "2025-01-15T12:01:30Z",
    "completed_at": null
  },
  "timestamp": "2025-01-15T12:01:30+00:00"
}
```

> 响应为裸 `Crawl` 域模型(无 `task` / `progress` 包装)。`status` 取 `CrawlStatus` 序列化值(`queued` / `processing` / `completed` / `failed` / `cancelled`)。

#### Get Crawl Results

**Endpoint:** `GET /v1/crawl/{id}/results`

**Parameters:**
- `id` (path) - Crawl UUID

> **无查询参数**:该端点不接受 `page` / `limit`,一次返回该 crawl 下全部子任务的抓取结果。

**Response:**
```json
{
  "success": true,
  "data": [
    {
      "id": "770e8400-e29b-41d4-a716-446655440000",
      "task_id": "880e8400-e29b-41d4-a716-446655440001",
      "url": "https://example.com/page1",
      "status_code": 200,
      "content": "<html>...</html>",
      "content_type": "text/html",
      "headers": {},
      "meta_data": {},
      "screenshot": null,
      "response_time_ms": 1200,
      "created_at": "2025-01-15T12:00:05Z"
    }
  ],
  "timestamp": "2025-01-15T12:00:05+00:00"
}
```

> 响应 `data` 是裸 `Vec<ScrapeResult>` 数组(非分页包封)。

#### Cancel Crawl

Cancel a crawl task. All sub-tasks of the crawl are cancelled with it.

**Endpoint:** `DELETE /v1/crawl/{id}`

**Parameters:**
- `id` (path) - Crawl UUID

**Response:** `204 No Content` — 空 body,无 JSON 包封(`errors` 时返回标准错误包封:403 非属主 / 404 不存在 / 500 内部错误)。

---

### Search API

Search using various search engines.

#### Search

**Endpoint:** `POST /v1/search`

**Request Body:**
```json
{
  "engine": "smart",
  "query": "Rust web scraping",
  "limit": 10,
  "lang": "en",
  "country": "us",
  "sync_wait_ms": 5000
}
```

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `engine` | string | No | Search engine: `smart`(默认,多引擎 fan-out + RRF 聚合)、`bing`、`baidu`、`sogou`。未提供或 `smart` 时并发查询多个引擎并做融合排序与去重;显式引擎名走单引擎路径;未知引擎名返回 400 参数错误(不会静默替换)。`exa`/`tavily` 保留枚举值,需对应引擎注册才可用,未注册时返回 `NoEngineAvailable` |
| `query` | string | Yes | Search query |
| `sources` | array | No | `engine` 的别名(multi-source 写法) |
| `limit` | integer | No | Number of results (default: 10; silently capped at `settings.search.max_results`, default **50**) |
| `lang` | string | No | Search language — **当前版本接受但尚未生效(已知缺口)**:`SearchService::perform_search` 的 `_lang` 参数被忽略;默认 `en` 的 Bing mkt/Accept-Language 驱动尚未实现(baidu/sogou 路径亦不读取该值) |
| `country` | string | No | Search region — **当前版本接受但尚未生效(已知缺口)**:`_country` 参数被忽略,默认 `us` 亦未驱动任何请求参数 |
| `sync_wait_ms` | integer | No | 同步等待时长(毫秒,默认 5000)。**注意:** 无 DTO 级上限校验(不同于 scrape/crawl 的显式 `MAX_SYNC_WAIT_MS` 检查),超过 30000 亦照常进入等待轮询 |
| `crawl_results` | boolean | No | 为 true 时对每条搜索结果创建爬取任务 |
| `crawl_config` | object | No | 搜索结果异步爬取的配置(结构同爬取任务配置) |

> **0.2.0 行为变更**:`engine` 省略时从"单引擎 bing"改为 smart 聚合(fan-out + RRF + 跨来源去重);
> 显式指定 `engine: "bing"` 等仍为单引擎路径。

**Response (Success):**
```json
{
  "success": true,
  "data": {
    "query": "Rust web scraping",
    "results": [
      {
        "title": "Web Scraping with Rust",
        "url": "https://example.com/rust-scraping",
        "description": "Learn how to scrape websites using Rust",
        "engine": "bing"
      }
    ],
    "crawl_id": "550e8400-e29b-41d4-a716-446655440000",
    "credits_used": 5
  }
}
```

> **`score` 字段(0.2.0 新增,可选)**:启用 `[rag]` 语义重排后,每条结果追加
> `score`(0..1 的 query-doc 相关性得分)。未启用时该字段整体省略
> (`skip_serializing_if = "Option::is_none"`),既有消费者的解析不受影响。

---

### Map API (new in 0.2.0)

Discover URLs from a site s's sitemap. Synchronous endpoint.

**Endpoint:** `POST /v1/map`

**Request Body:**
```json
{
  "url": "https://example.com",
  "include_patterns": ["*/blog/*"],
  "exclude_patterns": ["*/tag/*"],
  "limit": 1000
}
```

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `url` | string | Yes | Site origin; `{origin}/sitemap.xml` is fetched |
| `include_patterns` | string[] | No | Glob whitelist (`*`/`?`) applied first |
| `exclude_patterns` | string[] | No | Glob blacklist applied after include |
| `limit` | integer | No | Max URLs returned (default: 1000). No DTO-level cap — the overall total is bounded by `sitemap.max_urls` (default **50000**, `config/default.toml`) |

**Response:**
```json
{
  "success": true,
  "data": {
    "links": ["https://example.com/", "https://example.com/blog/post-1"]
  }
}
```

**Behavior notes:**
- Missing sitemap (404) returns 200 with empty `links` (valid state, not an error).
- Sitemap index (`<sitemapindex>`) is expanded with sub-sitemap fetches bounded by `max_concurrent_sub_sitemaps = 4` (`config/default.toml`), one level deep.

**Errors:** `400` invalid URL/params (`OriginUnparsable` / internal validation) · `502` `MAP_TARGET_UNREACHABLE` target unreachable or 5xx · unknown DTO fields rejected with `422` via `deny_unknown_fields`

---

### Extract API

> **Conditional Endpoint (R-teams-003):** This endpoint's signature depends on the `teams` feature:
> - **`teams` enabled (default):** `POST /v1/extract` accepts `Extension<Arc<TeamService>>` and `Extension<Arc<GR>>` (geo-restriction repository) parameters, enforcing team-scoped geo-restrictions before task creation.
> - **`teams` disabled:** `POST /v1/extract` degrades to a simpler signature without geo-restriction parameters; tasks are created directly against `DEFAULT_TEAM_ID` without geo checks.

Extract structured data from target pages. The endpoint creates an asynchronous extract task (it does **not** extract from inline HTML); results are fetched via the task status endpoints.

#### Extract Data

**Endpoint:** `POST /v1/extract`

**Request Body:**
```json
{
  "urls": ["https://example.com"],
  "prompt": "Extract the page title and all article links",
  "sync_wait_ms": 5000
}
```

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `urls` | string[] | Yes | Target URLs (1–100 entries, http/https only) |
| `prompt` (alias: `extraction_prompt`) | string | One of `prompt`/`schema`/`rules` required | Free-text LLM prompt describing what to extract |
| `schema` (alias: `extraction_schema`) | object | One of `prompt`/`schema`/`rules` required | JSON Schema for structured LLM extraction |
| `rules` (alias: `extraction_rules`) | object | One of `prompt`/`schema`/`rules` required | CSS selector extraction rules |
| `model` | string | No | Model identifier for LLM extraction |
| `sync_wait_ms` | integer | No | Synchronous wait in milliseconds (default 5000, max 30000; `> 30000` returns 400) |

Each `rules` entry (`ExtractionRule`) supports: `selector`, `attr` (attribute name; `None` extracts text), `is_array` (multiple values), `use_llm`, `llm_prompt`, `output_format` (`json` default / `plaintext`).

**Response (Success):**
```json
{
  "success": true,
  "data": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "status": "pending"
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

**Status codes:**

| Status | Meaning |
|--------|---------|
| `201` | Task enqueued, no `sync_wait_ms` (or 0) |
| `202` | `sync_wait_ms` elapsed with the task still not finished — the body still reports `status: "pending"` |
| `400` | `urls` empty, or none of `prompt`/`schema`/`rules` provided, or `sync_wait_ms > 30000` |
| `402` | Insufficient credits (`check_and_deduct_quota` failed; the task is not enqueued) |
| `403`/`500` | Geographic restriction denied / other server-side failure |

---

### Task API

Query and manage tasks. Task API follows RESTful conventions with action suffixes (`_query` for queries, `_cancel` for cancellations). Both endpoints are registered in the production router (`src/presentation/routes/task.rs::task_routes`, merged via `src/bootstrap/routes.rs::create_v2_routes_with_state`).

#### Query Tasks

**Endpoint:** `POST /v1/tasks/_query`

**Request Body:**
```json
{
  "task_ids": ["550e8400-e29b-41d4-a716-446655440000"],
  "task_types": ["scrape", "crawl", "extract"],
  "statuses": ["queued", "active", "completed", "failed", "cancelled"],
  "created_after": "2025-01-01T00:00:00Z",
  "created_before": "2025-01-15T00:00:00Z",
  "crawl_id": "770e8400-e29b-41d4-a716-446655440000",
  "limit": 20,
  "offset": 0,
  "include_results": false,
  "sync_wait_ms": 5000
}
```

> All fields are optional and flat (no `filters`/`pagination`/`sort` wrappers). Defaults: `limit` 100, `offset` 0, `include_results` false, `sync_wait_ms` 5000. `limit` range 1–1000 (0 or >1000 rejected with 422). `sync_wait_ms` range 0–30000.

**Parameters:**

| Parameter | Type | Description |
|-----------|------|-------------|
| `task_ids` | array | Query specific task IDs |
| `task_types` (plural) | array | Filter by type: `scrape`, `crawl`, `extract` |
| `statuses` | array | Filter by status: `queued`, `active`, `completed`, `failed`, `cancelled` |
| `created_after` | string | Filter by creation date (RFC3339 with offset) |
| `created_before` | string | Filter by creation date (RFC3339 with offset) |
| `crawl_id` | string | All sub-tasks of a crawl |
| `limit` | integer | Results per page (default 100, max 1000) |
| `offset` | integer | Pagination offset (0-based; no `page` parameter) |
| `include_results` | boolean | Include task result payload |
| `sync_wait_ms` | integer | Synchronous poll wait (default 5000, max 30000) |

**Response (Success):**
```json
{
  "success": true,
  "data": {
    "tasks": [
      {
        "id": "550e8400-e29b-41d4-a716-446655440000",
        "task_type": "scrape",
        "status": "completed",
        "priority": 0,
        "url": "https://example.com",
        "attempt_count": 1,
        "max_retries": 3,
        "created_at": "2025-01-15T12:00:00+00:00",
        "started_at": "2025-01-15T12:00:01+00:00",
        "completed_at": "2025-01-15T12:00:03+00:00",
        "crawl_id": null,
        "result": null
      }
    ],
    "total": 150,
    "has_more": true
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> 分页用 `limit`/`offset`:`has_more = offset + len(tasks) < total`。无 `page` 参数、无 `pagination` 包封。`status` 序列化为 `queued`/`active`/`completed`/`failed`/`cancelled`。

#### Cancel Tasks

**Endpoint:** `POST /v1/tasks/_cancel`

**Request Body:**
```json
{
  "task_ids": [
    "550e8400-e29b-41d4-a716-446655440000",
    "660e8400-e29b-41d4-a716-446655440001"
  ],
  "force": false,
  "sync_wait_ms": 5000
}
```

> `task_ids` is required and must be non-empty (empty list → 422). `force` (default false) cancels tasks even while they are executing. `sync_wait_ms` (default 5000, max 30000) polls until cancellation settles.

**Response:**
```json
{
  "success": true,
  "data": {
    "cancelled_tasks": [
      {
        "task_id": "550e8400-e29b-41d4-a716-446655440000",
        "status": "cancelled",
        "cancelled_at": "2025-01-15T12:00:05+00:00"
      }
    ],
    "failed_tasks": [
      {
        "task_id": "660e8400-e29b-41d4-a716-446655440001",
        "reason": "Already completed"
      }
    ],
    "total_cancelled": 1,
    "total_failed": 1
  },
  "timestamp": "2025-01-15T12:00:05+00:00"
}
```

---

### Team API

> **Conditional Endpoints (R-teams-002):** All `/v1/teams/*` endpoints are only available when the `teams` feature is enabled (default). When `teams` is disabled, these routes are not registered and return 404 Not Found. In single-tenant mode (`teams` disabled), all requests are attributed to `DEFAULT_TEAM_ID` (`domain::id::Id::from_u128(1)`), making team management endpoints unnecessary.

#### Get Current Team

**Endpoint:** `GET /v1/teams/me`

**Response:**
```json
{
  "success": true,
  "data": {
    "id": "770e8400-e29b-41d4-a716-446655440000",
    "name": "Team",
    "credits_balance": 1000,
    "total_tasks": 50,
    "completed_tasks": 0,
    "failed_tasks": 0,
    "created_at": "2025-01-15T12:00:00Z"
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> `total_tasks` counts tasks found via `find_by_crawl_id(team_id)`; `completed_tasks`/`failed_tasks` are currently fixed `0`(handler 统计尚未接通)。`credits_balance` 查询失败时回退为 `0`(不报错)。

#### Get Team Usage

**Endpoint:** `GET /v1/teams/me/usage`

**Response:**
```json
{
  "success": true,
  "data": {
    "team_id": "770e8400-e29b-41d4-a716-446655440000",
    "period": "30d",
    "total_requests": 0,
    "successful_requests": 0,
    "failed_requests": 0,
    "credits_used": 1234,
    "avg_response_time_ms": 123.45
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> `period` 固定 `"30d"`。`total_requests`/`successful_requests`/`failed_requests` 当前恒为 `0`(handler 统计尚未接通);`credits_used` 取积分余额绝对值(负余额表示超支)。无 `credits_limit` / `requests_today` / `requests_limit` 字段。`avg_response_time_ms` 为该团队抓取结果平均响应时间。

#### Get Team Geo Restrictions

**Endpoint:** `GET /v1/teams/geo-restrictions`

**Response:**
```json
{
  "success": true,
  "data": {
    "team_id": "770e8400-e29b-41d4-a716-446655440000",
    "enable_geo_restrictions": true,
    "allowed_countries": ["US", "CA"],
    "blocked_countries": ["CN", "RU"],
    "ip_whitelist": ["10.0.0.0/8"],
    "domain_blacklist": ["evil.com"]
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> 响应为平铺的 `TeamGeoRestrictionsResponse`(无 `restrictions` 外层包装)。字段名是 `enable_geo_restrictions`(不是 `enabled`);未设置的列表字段序列化为 `null`。

#### Update Team Geo Restrictions

**Endpoint:** `PUT /v1/teams/geo-restrictions`

**Request Body:**
```json
{
  "enable_geo_restrictions": true,
  "allowed_countries": ["US", "CA"],
  "blocked_countries": ["CN", "RU"],
  "ip_whitelist": ["10.0.0.0/8"],
  "domain_blacklist": ["evil.com"]
}
```

> `enable_geo_restrictions` **必填**(bool,无缺省值)。国家代码必须为 2 字母 ISO 3166-1 alpha-2,`ip_whitelist` 支持 IP 或 CIDR——校验失败返回 `400 VALIDATION_ERROR`。`deny_unknown_fields` 拒绝未知字段(422)。allowed/blocked 列表若提供则不能为空数组。

**Response:**
```json
{
  "success": true,
  "data": {
    "team_id": "770e8400-e29b-41d4-a716-446655440000",
    "enable_geo_restrictions": true,
    "allowed_countries": ["US", "CA"],
    "blocked_countries": ["CN", "RU"],
    "ip_whitelist": ["10.0.0.0/8"],
    "domain_blacklist": ["evil.com"]
  },
  "timestamp": "2025-01-15T12:00:00+00:00"
}
```

> 响应为更新后的 restrictions 对象本身(非 `{message}` 包封)。

---

### Webhook API

> **Conditional Endpoints (R-wh-001):** All `/v1/webhooks/*` endpoints are only available when the `webhook` feature is enabled (default). When `webhook` is disabled, these routes are not registered and return 404 Not Found. Additionally, the `webhook_worker` is not spawned, and `NoopWebhookService` is injected (task completion/failure notifications become no-ops).

#### List Webhooks

**Endpoint:** `GET /v1/webhooks`

> **无查询参数**:一次返回该团队全部 webhook(无 `page` / `limit`)。列表响应**不含** `secret` 字段。

**Response:**
```json
{
  "success": true,
  "data": {
    "webhooks": [
      {
        "id": "550e8400-e29b-41d4-a716-446655440000",
        "team_id": "660e8400-e29b-41d4-a716-446655440001",
        "url": "https://your-webhook.com/callback",
        "created_at": "2026-01-15T00:00:00Z",
        "is_active": true
      }
    ],
    "total": 5
  },
  "timestamp": "2026-01-15T00:00:00+00:00"
}
```

#### Create Webhook

**Endpoint:** `POST /v1/webhooks`

**Request Body:**
```json
{
  "url": "https://your-webhook.com/callback"
}
```

> 请求体经 `#[serde(deny_unknown_fields)]` 校验:`CreateWebhookRequest` 仅接受 `url`,传入 `events` / `secret` / `active` 等未知字段会被拒绝并返回 **400 `VALIDATION_ERROR`**(handler 以 `serde_json::from_slice` 解析后映射为 `CrawlRsError::Validation`)。此外,创建请求本身必须携带按服务端全局密钥(`CRAWLRS__WEBHOOK__SECRET`)计算的 Standard Webhooks 签名头(见 [Webhook Signature](#webhook-signature)),签名验证失败返回 **401 `AUTHENTICATION_ERROR`**。

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `url` | string | Yes | Webhook callback URL (http/https) |

**Events(服务端自动按任务类型派发,无需订阅):**
- `scrape.completed` - Scrape task completed
- `scrape.failed` - Scrape task failed
- `crawl.completed` - Crawl task completed
- `crawl.failed` - Crawl task failed
- `extract.completed` / `extract.failed` - Extract task completed/failed

**Response (Success):** `201 Created`,标准 `{"success":true,"data":{...}}` 包封。**创建响应返回 `secret`**(各 webhook 的签名密钥由服务端按 `settings.webhook.secret` 统一派生,仅在本次创建响应中返回;列表端点不回传)。

```json
{
  "success": true,
  "data": {
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "team_id": "660e8400-e29b-41d4-a716-446655440001",
    "url": "https://your-webhook.com/callback",
    "created_at": "2026-01-15T00:00:00Z",
    "is_active": true,
    "secret": "your-webhook-signing-secret"
  },
  "timestamp": "2026-01-15T00:00:00+00:00"
}
```

---

### Audit API

#### Get Audit Logs

**Endpoint:** `GET /v1/audit/logs`

**Query Parameters:**
- `limit` - Results per page (default `server_config::DEFAULT_PAGE_LIMIT`, capped at `MAX_PAGE_LIMIT`)
- `offset` - Pagination offset (default 0)
- `api_key_id` - Filter by API key (non-admin callers may only pass their own key; otherwise 403)
- `team_id` - Filter by team (non-admin callers may only pass their own team; otherwise 403)

> 无 `event_type` / `start_time` / `end_time` / `page` 参数。缺省查询调用方自身 `api_key_id` 的日志;admin scope 可查任意 key/team。IDOR 防护见 `audit_handler.rs`。

**Response:**
```json
{
  "success": true,
  "data": {
    "logs": [
      {
        "id": "550e8400-e29b-41d4-a716-446655440000",
        "api_key_id": "770e8400-e29b-41d4-a716-446655440000",
        "team_id": "660e8400-e29b-41d4-a716-446655440001",
        "requested_action": "GET /v1/scrape",
        "decision": "Allow",
        "denial_reason": null,
        "scope_used": null,
        "ip_address": "192.168.1.1",
        "trace_id": null,
        "user_agent": "Mozilla/5.0...",
        "request_path": "/v1/scrape",
        "request_method": "GET",
        "metadata": {},
        "created_at": "2025-01-15T00:00:00Z"
      }
    ]
  },
  "timestamp": "2025-01-15T00:00:00+00:00"
}
```

> 响应仅含 `logs` 数组(`AuditLogEntry`),无 `pagination` 包封。`decision` 序列化为 `"Allow"` / `"Deny"`。

#### Get Denied Requests

**Endpoint:** `GET /v1/audit/denied`

**Query Parameters:**
- `limit` - Results per page (default `server_config::DEFAULT_PAGE_LIMIT`, capped at `MAX_PAGE_LIMIT`)

> 仅支持 `limit`;不支持 `reason` / `start_time` / `end_time` / `page`。始终查询调用方自身 `api_key_id` 的拒绝记录(避免 IDOR)。

**Response:**
```json
{
  "success": true,
  "data": {
    "denied_requests": [
      {
        "id": "550e8400-e29b-41d4-a716-446655440000",
        "api_key_id": "770e8400-e29b-41d4-a716-446655440000",
        "team_id": "660e8400-e29b-41d4-a716-446655440001",
        "requested_action": "POST /v1/scrape",
        "decision": "Deny",
        "denial_reason": "rate_limit_exceeded",
        "scope_used": null,
        "ip_address": "192.168.1.1",
        "trace_id": null,
        "user_agent": "Mozilla/5.0...",
        "request_path": "/v1/scrape",
        "request_method": "POST",
        "metadata": {},
        "created_at": "2025-01-15T00:00:00Z"
      }
    ]
  },
  "timestamp": "2025-01-15T00:00:00+00:00"
}
```

> 响应仅含 `denied_requests` 数组(同为 `AuditLogEntry`,`decision` 为 `"Deny"`),无 `pagination` 包封。

---

### Admin API

> **0.2.0 新增(`garrison-auth-migration`):** 管理员可通过此端点签发新的 garrison API Key。要求调用方持 `crawlrs:admin` 权限(即 `admin` scope)。

#### Issue API Key

为指定 team 签发新的 garrison API Key。明文 key 仅返回一次,响应头含 `Cache-Control: no-store` / `Pragma: no-cache` / `Expires: 0` 防止缓存。

**Endpoint:** `POST /v1/admin/api-keys`

**Required Scope:** `admin`(对应 garrison 权限 `crawlrs:admin`)

**Request Body:**
```json
{
  "team_id": "770e8400-e29b-41d4-a716-446655440000",
  "scopes": ["read", "write", "admin"],
  "expires_in_secs": 2592000
}
```

**Parameters:**

| Parameter | Type | Required | Description |
|-----------|-------|----------|-------------|
| `team_id` | string (UUID) | Yes | 该 API Key 所属的 team ID |
| `scopes` | array<string> | Yes | scope 列表,可选值:`read` / `write` / `admin`(handler 自动映射为 garrison 权限 `crawlrs:read` / `crawlrs:write` / `crawlrs:admin`) |
| `expires_in_secs` | integer | No | 过期秒数(默认 2592000 = 30 天) |

**Request Example:**
```bash
curl -X POST http://localhost:8899/v1/admin/api-keys \
  -H "Authorization: Bearer garrison_admin_key_id.garrison_admin_secret" \
  -H "Content-Type: application/json" \
  -d '{
    "team_id": "770e8400-e29b-41d4-a716-446655440000",
    "scopes": ["read", "write"],
    "expires_in_secs": 2592000
  }'
```

**Response Headers:**
```
Cache-Control: no-store
Pragma: no-cache
Expires: 0
```

**Response (201 Created):**
```json
{
  "success": true,
  "data": {
    "api_key": "garrison_key_id.garrison_secret",
    "api_key_id": "880e8400-e29b-41d4-a716-446655440000",
    "team_id": "770e8400-e29b-41d4-a716-446655440000",
    "scopes": ["read", "write"],
    "expires_at": "2025-08-20T12:00:00+00:00"
  },
  "timestamp": "2025-07-21T12:00:00+00:00"
}
```

> **注:** 响应中 `scopes` 字段回显请求时传入的 scope(`read`/`write`/`admin`),实际 garrison 权限串(`crawlrs:read` 等)由 handler 在签发时映射,调用方无需关心。`expires_at` 为 RFC3339 过期时间(缺省 30 天;`expires_in_secs` 必须 > 0 且 ≤ 1 年,`scopes` 最多 10 个元素)。

> **⚠️ 安全提示:** `api_key` 字段是明文 key,仅在本次响应中返回一次。请立即妥善保存(如写入 secrets manager),后续无法再次查询。若丢失需重新签发。

**Error Codes:**

| Error Code | HTTP Status | Description |
|------------|-------------|-------------|
| `UNAUTHORIZED` | 401 | 调用方未认证或 garrison key 无效 |
| `FORBIDDEN` | 403 | 调用方缺少 `crawlrs:admin` 权限(handler 内 `has_permission(Admin)` 校验) |
| `VALIDATION_ERROR` | 400 | 请求值校验失败(`team_id` 为 nil UUID、`scopes` 为空/含未知权限、`expires_in_secs` 越界) |
| `UNPROCESSABLE_ENTITY` | 422 | 请求体结构错误(如 `team_id` 非 UUID 的 JSON 数据错误、字段缺失/类型不符) |
| `INTERNAL_ERROR` | 500 | `team_id` 在 teams 表不存在、garrison DAO 未注入或 `ApiKeyHandler::generate_with_namespace` 内部错误 |

> **注意:** `auth` feature 未启用时路由不注册,请求得到普通 404(`NOT_FOUND`),不存在 `FEATURE_DISABLED` 响应码。

---

## 🎭 Session Identity & Cookie Semantics

When `[session].enabled = true` (default: `false`), the engine router binds consecutive
requests to a session identity (proxy + UA/TLS profile + cookie cache).

- **Cookie scope**: response `Set-Cookie` values are captured per target host and only
  replayed to requests targeting the **same host**. Cookies captured from site A are never
  injected into requests to site B, even when requests share one `session_id` across hosts.
- **Known limitations of the manual cookie channel**: no `expiry`/`path` attribute handling
  (short-lived cookies may be replayed past expiration; only the first `name=value` pair of
  each `Set-Cookie` is kept). Disable `[session].enabled` if the target site relies on
  strict cookie semantics. Browser engines (Playwright/MLLM) manage cookies in their own
  context and do not use this channel.

### XHR Capture (`capture_xhr`)

When `CRAWLRS__ENGINES__PLAYWRIGHT__CAPTURE_XHR=true`, the Playwright engine can capture
background XHR/Fetch responses:

- Captured data is persisted with the task result under `meta_data.captured_responses`
  (read it via the task result endpoint). The `POST /v1/scrape` enqueue response itself
  does not carry capture data (`captured_responses` is always `null` there).
- XHR/Fetch URLs pointing to loopback/private/link-local addresses are discarded (SSRF
  protection). Configure `capture_xhr_url_patterns` as a whitelist to limit what gets
  captured.

---

## ⏱️ Rate Limiting

> **Conditional Behavior (R-rl-003):** Rate limiting is only active when the `rate-limit` feature is enabled (default). When disabled, `NoopRateLimitingService` is injected: `check_rate_limit` returns `Allowed`, `check_and_deduct_quota` returns `Ok(())`, `get_quota_balance` returns `Ok(i64::MAX)`. All requests are allowed without limit, and quota deductions are no-ops.

The API implements rate limiting at multiple levels:

1. **Per-API Key Rate Limit** - Limits requests per API key
2. **Per-Team Concurrency Limit** - Limits concurrent requests per team
3. **Global Rate Limit** - System-wide protection

### Rate Limit Headers

**当前实现仅设一个响应头:** 命中限流返回 429 时设 `Retry-After`。**不存在 `X-RateLimit-Limit` / `X-RateLimit-Remaining` / `X-RateLimit-Reset` 响应头**(全库无发射点,读取这些头的客户端逻辑应移除)。

```
Retry-After: 60
```

响应体示例见上方 [Rate Limit Error Response](#rate-limit-error-response):标准 `ApiResponse` 包封(`error.code = "RATE_LIMITED"`),`RetryAfter` 结果额外带 `retry_after_seconds`。

> 附加语义:限流服务本身出错时默认 fail-closed(返回 503 `SERVICE_UNAVAILABLE`);设 `RATE_LIMIT_FAIL_OPEN=true` 可改为放行并记录 warn 日志。

---

## 🪝 Webhooks

Webhooks allow you to receive notifications about task events.

### Webhook Payload Format

投递的 payload 为包含 `task_id`、`status`、`url`、`timestamp`(Unix 秒)的 JSON 对象,失败时附带 `error`:

```json
{
  "task_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "completed",
  "url": "https://example.com",
  "timestamp": 1760000000
}
```

事件类型由任务类型决定(`scrape.completed` / `scrape.failed` / `crawl.completed` / `crawl.failed` / `extract.completed` / `extract.failed`)。

### Webhook Signature

签名遵循 [Standard Webhooks](https://www.standardwebhooks.com/) 规范,投递携带三个头:

```
webhook-id: <msg_id>
webhook-timestamp: <unix_seconds>
webhook-signature: v1,<base64-encoded-hmac-sha256>
```

签名消息为 `{msg_id}.{timestamp}.{payload}`(三段以句点拼接),使用服务端配置的密钥(`CRAWLRS__WEBHOOK__SECRET`)计算 HMAC-SHA256。验证时用 `webhook-id` 与 `webhook-timestamp` 头的值重算并对比 `webhook-signature`(推荐使用官方 `standardwebhooks` 库,恒时比较)。

---

## 🧩 SDK API

SDK endpoints provide simplified interfaces for common operations, wrapping the underlying REST API. All SDK endpoints return the payload object directly (no `success`/`data` envelope).

> **路径解析:** SDK 路由由 sdforge `#[forge]` 宏注册(`version = "v1"` + `path = "/sdk/..."`),sdforge 默认前缀为 `/api/{version}`,因此实际对外路径统一为 `/api/v1/sdk/*`。SDK 路由同样经过认证中间件(`team_id`/`api_key_id` 从 `AuthState` 提取,绝不信任请求体)。

### SDK Search

**Endpoint:** `POST /api/v1/sdk/search`

**Request Body:**
```json
{
  "query": "Rust web scraping",
  "limit": 10
}
```

| Parameter | Type | Required | Description |
|-----------|------|----------|-------------|
| `query` | string | Yes | Search query (non-empty) |
| `limit` | integer | No | Max results |

**Response:**
```json
{
  "query": "Rust web scraping",
  "results": [
    {
      "title": "...",
      "url": "https://...",
      "description": "...",
      "engine": "bing"
    }
  ],
  "credits_used": 5
}
```

### SDK Tasks

Create a queued task (`task_type`: `scrape` or `crawl`).

**Endpoint:** `POST /api/v1/sdk/tasks`

**Request Body:**
```json
{
  "url": "https://example.com",
  "task_type": "scrape"
}
```

**Response:**
```json
{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "url": "https://example.com",
  "status": "Queued"
}
```

### SDK Scrape

**Endpoint:** `POST /api/v1/sdk/scrape`

**Request Body:**
```json
{
  "url": "https://example.com"
}
```

**Response:**
```json
{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "url": "https://example.com"
}
```

### SDK Crawl

**Endpoint:** `POST /api/v1/sdk/crawl`

**Request Body:**
```json
{
  "name": "my-crawl",
  "url": "https://example.com",
  "seed_url": "https://example.com",
  "max_depth": 2
}
```

`max_depth` 可选,缺省 2,范围 0–5(与 HTTP crawl 入口同源上限)。
创建时即为 `seed_url` 播种 depth=0 的首任务,`total_tasks` 从 1 起算;
首任务入队失败时接口返回 500,并把该 crawl 置为 Failed(计数自洽,不留孤儿)。
与 `/v1/crawl` 一致:`url`/`seed_url` 仅接受 http(s),并执行团队 `domain_blacklist`
(命中返回 400,黑名单随 crawl.config 持久化供 worker 与链接过滤使用);地理(国家)限制依赖请求源 IP,
服务端到服务端调用无客户端 IP,因此 SDK 入口不做国家判定。
SDK 端点按 API key 做限流(`/sdk/crawl`),该入口不计 credit(与 `/sdk/search`
的按次计费口径不同,`/sdk/crawl` 的额度由限流约束)。

**Response:**
```json
{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "Queued",
  "url": "https://example.com"
}
```

---

## 💻 SDK Examples

### JavaScript/Node.js

```javascript
const axios = require('axios');

const client = axios.create({
  baseURL: 'http://localhost:8899',
  headers: {
    'Authorization': 'Bearer YOUR_API_KEY'
  }
});

const scrape = async (url) => {
  const response = await client.post('/v1/scrape', {
    url: url,
    formats: ['markdown'],
    sync_wait_ms: 5000
  });
  return response.data;
};
```

### Python

```python
import requests

client = requests.Session()
client.headers.update({
  'Authorization': 'Bearer YOUR_API_KEY'
})

def scrape(url):
  response = client.post('http://localhost:8899/v1/scrape', json={
    'url': url,
    'formats': ['markdown'],
    'sync_wait_ms': 5000
  })
  return response.json()
```

### Go

```go
package main

import (
  "bytes"
  "encoding/json"
  "net/http"
)

func Scrape(url string) error {
  client := &http.Client{}
  body := map[string]interface{}{
    "url": url,
    "formats": []string{"markdown"},
    "sync_wait_ms": 5000,
  }

  jsonData, _ := json.Marshal(body)
  req, _ := http.NewRequest("POST", "http://localhost:8899/v1/scrape", bytes.NewBuffer(jsonData))
  req.Header.Set("Authorization", "Bearer YOUR_API_KEY")
  req.Header.Set("Content-Type", "application/json")

  resp, err := client.Do(req)
  return err
}
```

---

## ✅ Best Practices

1. **Use Sync Mode Sparingly** - Only use `sync_wait_ms` when you need immediate results
2. **Implement Retry Logic** - Handle rate limits with exponential backoff
3. **Use Webhooks** - Prefer webhooks over polling for task status
4. **Set Timeouts** - Always configure appropriate timeouts
5. **Monitor Credits** - Track credit usage to avoid service interruption
6. **Handle Errors Gracefully** - Check both HTTP status and response `success` field
7. **Validate Inputs** - Validate URLs and parameters before sending requests
8. **Use Caching** - Enable oxcache for frequently accessed content
9. **Set Proper Rates** - Configure rate limits appropriate for your capacity
10. **Secure Webhooks** - Always verify webhook signatures

---

## 📋 Changelog

### Unreleased

**Changed:**
- ID 体系全量迁移至 NebulaId(`domain::id::Id`,默认 `uuid_v8` 算法;API JSON 契约保持 UUID 字符串不变)
- `/v1/search` 无 `engine` 参数时行为从"单引擎 bing"变为 smart 聚合(多引擎 fan-out + RRF 融合)
- auth-off 语义:protected routes 默认 401 拒绝,须显式 `allow_unauthenticated_protected()` opt-in;opt-in 后注入 `ApiKeyScope::denied()` 受限匿名身份(`DEFAULT_TEAM_ID`/`DEFAULT_API_KEY_ID`),非 `full_access()`

**New Engines / Algorithms:**
- TLS Fingerprint Engine (`engine-tls-fingerprint`): WreqEngine based on BoringSSL, real JA3/JA4 fingerprint spoofing
- MLLM Autonomous Navigation Engine (`engine-mllm`): Vision LLM agentic loop for autonomous page navigation
- RAG 增强提取、知识图谱覆盖感知爬取、DRL 自适应爬取策略

**Observability:**
- 5 个新 Prometheus 指标:`crawlrs_queue_depth` / `crawlrs_engine_success_total` / `crawlrs_engine_duration_seconds` / `crawlrs_cache_hit_total` / `crawlrs_webhook_delivery_total`

### 0.2.0 (2026-11-15)
- 首个 crates.io 稳定版:自研依赖归一稳定版(minor 精度)、alchemize 切 registry(无 path/git 依赖)、i18n 宿主注册重写、release.yml OIDC + token 兜底
- acceptance teardown clippy 修复;SDK 测试 I18nBundle 两参调用

### 0.2.0-rc.4 (2026-11-10)
- 发布流水线全量补跑:CI 五组合/engine-moli/BDD 修复复绿;CodeQL 18 条裁决归零;security 审计豁免同源登记
- doctest/示例/文档一致性修复;依赖 html-escape 0.3.1 / toml 1.2.1 / uuid 1.29.2
- 行覆盖率 90.13%,E2E matrix/static/unit/integration/bench/liveapi 六阶段 PASS(report 汇总按设计 SKIPPED)

### 0.2.0-rc.3 (2026-11-10)
- crawl 终态判定单源 + 崩溃恢复计数重算;团队域名黑名单执行链路(种子/重定向/链接/sitemap/proxy,快照缺失 fail-closed)
- SDK `POST /api/v1/sdk/crawl`;搜索派生 200 硬上限与深度 ≤5;scrape/crawl 统一丢弃自定义 Host 头
- release.yml CHANGELOG 路径对齐 docs/;git 依赖期间 crates.io 发布显式跳过并附 tag 源码归档

### 0.2.0-rc.2 (2026-09-21)
- 自研依赖全面切换 crates.io registry(garrison → 0.9.0-rc.2 等),移除跨仓 path 依赖

### 0.2.0-rc.1 (2026-08-25)
- auth-off 安全语义收紧:protected routes 默认 401,`allow_unauthenticated_protected()` 显式 opt-in(该收紧为现行语义,见上方 Auth Disabled Behavior 节)

### 0.2.0-legacy (2026-07-29)
- **`garrison-auth-migration`:** 认证引擎由 garrison 接管,Bearer token 格式改为 `garrison_key_id.garrison_secret`
- 新增 `POST /v1/admin/api-keys` 端点:管理员为指定 team 签发 garrison API Key
- 401/429 触发逻辑由 garrison `firewall-bruteforce` 接管(5 次/60 秒/300 秒锁定)
- Scrape API `options` 新增 `cache_mode`(4 种缓存读写模式:`enabled` / `disabled` / `read_only` / `bypass`)和 `bypass_cache`(应急缓存绕过)字段
- Added `GET /v1/webhooks` to list webhooks
- Added `GET /v1/teams/me` and `GET /v1/teams/me/usage` team endpoints
- Added SDK API section (`/api/v1/sdk/*` endpoints)
- Merged Task API sections into single unified section

### 0.1.0 (2026-07-22)
- Initial release
- Scrape, Crawl, Search, Extract APIs
- Rate limiting and concurrency control
- Webhook support
- Audit logging
- Metrics export

---

## 📞 Support

For questions or issues:
- 📖 [Documentation](../README.md)
- 🐛 [Issue Tracker](https://github.com/Kirky-X/crawlrs/issues)
- 📧 Email: Kirky-X@outlook.com