speccrew 0.1.12 → 0.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/.speccrew/agents/speccrew-feature-designer.md +120 -0
- package/.speccrew/agents/speccrew-product-manager.md +54 -0
- package/.speccrew/agents/speccrew-system-designer.md +150 -14
- package/.speccrew/agents/speccrew-system-developer.md +309 -37
- package/.speccrew/agents/speccrew-task-worker.md +43 -0
- package/.speccrew/agents/speccrew-team-leader.md +108 -11
- package/.speccrew/agents/speccrew-test-manager.md +278 -0
- package/.speccrew/skills/speccrew-dev-backend/SKILL.md +44 -0
- package/.speccrew/skills/speccrew-dev-desktop/SKILL.md +44 -0
- package/.speccrew/skills/speccrew-dev-frontend/SKILL.md +44 -0
- package/.speccrew/skills/speccrew-dev-mobile/SKILL.md +44 -0
- package/.speccrew/skills/speccrew-fd-api-contract/SKILL.md +70 -0
- package/.speccrew/skills/speccrew-fd-feature-design/SKILL.md +158 -0
- package/.speccrew/skills/speccrew-pm-requirement-analysis/SKILL.md +65 -0
- package/.speccrew/skills/speccrew-sd-backend/SKILL.md +38 -0
- package/.speccrew/skills/speccrew-sd-desktop/SKILL.md +38 -0
- package/.speccrew/skills/speccrew-sd-frontend/SKILL.md +38 -0
- package/.speccrew/skills/speccrew-sd-mobile/SKILL.md +38 -0
- package/.speccrew/skills/speccrew-test-case-design/SKILL.md +33 -0
- package/.speccrew/skills/speccrew-test-code-gen/SKILL.md +34 -0
- package/.speccrew/skills/speccrew-test-execute/SKILL.md +34 -0
- package/README.ar.md +70 -3
- package/README.bn.md +52 -0
- package/README.bs.md +70 -3
- package/README.da.md +70 -3
- package/README.de.md +70 -3
- package/README.el.md +52 -0
- package/README.en.md +69 -2
- package/README.es.md +70 -3
- package/README.fr.md +70 -3
- package/README.it.md +70 -3
- package/README.ja.md +70 -3
- package/README.ko.md +70 -3
- package/README.md +69 -2
- package/README.no.md +70 -3
- package/README.pl.md +70 -3
- package/README.pt-BR.md +70 -3
- package/README.ru.md +70 -3
- package/README.th.md +69 -2
- package/README.tr.md +69 -2
- package/README.uk.md +69 -2
- package/README.vi.md +52 -0
- package/README.zh-TW.md +70 -3
- package/docs/GETTING-STARTED.ar.md +78 -4
- package/docs/GETTING-STARTED.bn.md +78 -4
- package/docs/GETTING-STARTED.bs.md +78 -4
- package/docs/GETTING-STARTED.da.md +78 -4
- package/docs/GETTING-STARTED.de.md +78 -4
- package/docs/GETTING-STARTED.el.md +78 -4
- package/docs/GETTING-STARTED.en.md +78 -4
- package/docs/GETTING-STARTED.es.md +78 -4
- package/docs/GETTING-STARTED.fr.md +78 -4
- package/docs/GETTING-STARTED.it.md +78 -4
- package/docs/GETTING-STARTED.ja.md +79 -5
- package/docs/GETTING-STARTED.ko.md +79 -5
- package/docs/GETTING-STARTED.md +78 -4
- package/docs/GETTING-STARTED.no.md +78 -4
- package/docs/GETTING-STARTED.pl.md +78 -4
- package/docs/GETTING-STARTED.pt-BR.md +78 -4
- package/docs/GETTING-STARTED.ru.md +78 -4
- package/docs/GETTING-STARTED.th.md +79 -5
- package/docs/GETTING-STARTED.tr.md +78 -4
- package/docs/GETTING-STARTED.uk.md +78 -4
- package/docs/GETTING-STARTED.vi.md +79 -5
- package/docs/GETTING-STARTED.zh-TW.md +79 -5
- package/package.json +1 -1
- package/.speccrew/skills/speccrew-create-agents/SKILL.md +0 -98
- package/.speccrew/skills/speccrew-create-agents/templates/agents/designer-agent.md +0 -54
- package/.speccrew/skills/speccrew-create-agents/templates/agents/dev-agent.md +0 -79
- package/.speccrew/skills/speccrew-create-agents/templates/agents/test-agent.md +0 -80
- package/.speccrew/skills/speccrew-project-diagnosis/SKILL.md +0 -233
- package/.speccrew/skills/speccrew-project-diagnosis/templates/DIAGNOSIS-REPORT-TEMPLATE.md +0 -202
- package/.speccrew/skills/speccrew-workflow-diagnose/SKILL.md +0 -155
- package/workspace-template/docs/solutions/Agent/346/212/200/350/203/275/345/256/232/344/271/211+/351/234/200/346/261/202/346/226/207/346/241/243+UML/344/275/277/347/224/250/346/250/241/346/235/277/357/274/210ISA-95/345/205/255/346/256/265/345/274/217/350/236/215/345/220/210/347/211/210/357/274/211.md +0 -586
- package/workspace-template/docs/solutions/harness.md +0 -410
|
@@ -361,7 +361,67 @@ knowledges/techs/{platform-id}/
|
|
|
361
361
|
|
|
362
362
|
---
|
|
363
363
|
|
|
364
|
-
## 6.
|
|
364
|
+
## 6. การจัดการความคืบหน้าของ Pipeline
|
|
365
|
+
|
|
366
|
+
ทีมเสมือน SpecCrew ปฏิบัติตามกลไกการควบคุมขั้นตอนที่เข้มงวด แต่ละขั้นตอนต้องได้รับการยืนยันจากผู้ใช้ก่อนจึงจะสามารถดำเนินการต่อไปยังขั้นตอนถัดไป นอกจากนี้ยังรองรับการดำเนินการต่อจากจุดที่หยุด —— เมื่อรีสตาร์ทหลังจากหยุดชะงัก ระบบจะดำเนินการต่อโดยอัตโนมัติจากตำแหน่งที่หยุดไว้ครั้งล่าสุด
|
|
367
|
+
|
|
368
|
+
### 6.1 ไฟล์ความคืบหน้าสามระดับ
|
|
369
|
+
|
|
370
|
+
เวิร์กโฟลว์จะบำรุงรักษาไฟล์ความคืบหน้า JSON สามประเภทโดยอัตโนมัติ ตั้งอยู่ในไดเรกทอรีการทำซ้ำ:
|
|
371
|
+
|
|
372
|
+
| ไฟล์ | ตำแหน่ง | หน้าที่ |
|
|
373
|
+
|------|----------|--------|
|
|
374
|
+
| `WORKFLOW-PROGRESS.json` | `iterations/{iter}/` | บันทึกสถานะของแต่ละขั้นตอนใน Pipeline ทั้งหมด |
|
|
375
|
+
| `.checkpoints.json` | ภายใต้ไดเรกทอรีของแต่ละขั้นตอน | บันทึกสถานะการผ่านจุดยืนยัน (Checkpoint) ของผู้ใช้ |
|
|
376
|
+
| `DISPATCH-PROGRESS.json` | ภายใต้ไดเรกทอรีของแต่ละขั้นตอน | บันทึกความคืบหน้ารายการของงานขนาน (หลายแพลตฟอร์ม/โมดูล) |
|
|
377
|
+
|
|
378
|
+
### 6.2 การเปลี่ยนสถานะของขั้นตอน
|
|
379
|
+
|
|
380
|
+
แต่ละขั้นตอนปฏิบัติตามการเปลี่ยนสถานะดังต่อไปนี้:
|
|
381
|
+
|
|
382
|
+
```
|
|
383
|
+
pending → in_progress → completed → confirmed
|
|
384
|
+
```
|
|
385
|
+
|
|
386
|
+
- **pending**: ยังไม่เริ่มต้น
|
|
387
|
+
- **in_progress**: กำลังดำเนินการ
|
|
388
|
+
- **completed**: Agent ดำเนินการเสร็จสิ้น รอการยืนยันจากผู้ใช้
|
|
389
|
+
- **confirmed**: ผู้ใช้ยืนยัน Checkpoint สุดท้าย สามารถเริ่มขั้นตอนถัดไปได้
|
|
390
|
+
|
|
391
|
+
### 6.3 การดำเนินการต่อจากจุดที่หยุด
|
|
392
|
+
|
|
393
|
+
เมื่อรีสตาร์ท Agent ของขั้นตอนใดขั้นตอนหนึ่ง:
|
|
394
|
+
|
|
395
|
+
1. **ตรวจสอบ upstream โดยอัตโนมัติ**: ตรวจสอบว่าขั้นตอนก่อนหน้าได้รับการยืนยันแล้วหรือไม่ หากไม่ได้รับการยืนยันจะถูกบล็อกและแจ้งเตือน
|
|
396
|
+
2. **กู้คืน Checkpoint**: อ่าน `.checkpoints.json` ข้ามจุดยืนยันที่ผ่านไปแล้ว และดำเนินการต่อจากจุดที่หยุดครั้งล่าสุด
|
|
397
|
+
3. **กู้คืนงานขนาน**: อ่าน `DISPATCH-PROGRESS.json` รีเอ็กซ์คิวต์เฉพาะงานที่มีสถานะ `pending` หรือ `failed` ข้ามงานที่มีสถานะ `completed` แล้ว
|
|
398
|
+
|
|
399
|
+
### 6.4 ดูความคืบหน้าปัจจุบัน
|
|
400
|
+
|
|
401
|
+
ดูสถานะภาพรวมของ Pipeline ผ่าน Team Leader Agent:
|
|
402
|
+
|
|
403
|
+
```
|
|
404
|
+
@speccrew-team-leader ดูความคืบหน้าของการทำซ้ำปัจจุบัน
|
|
405
|
+
```
|
|
406
|
+
|
|
407
|
+
Team Leader จะอ่านไฟล์ความคืบหน้าและแสดงสรุปสถานะที่คล้ายกับต่อไปนี้:
|
|
408
|
+
|
|
409
|
+
```
|
|
410
|
+
Pipeline Status: i001-user-management
|
|
411
|
+
01 PRD: ✅ Confirmed
|
|
412
|
+
02 Feature Design: 🔄 In Progress (Checkpoint A passed)
|
|
413
|
+
03 System Design: ⏳ Pending
|
|
414
|
+
04 Development: ⏳ Pending
|
|
415
|
+
05 System Test: ⏳ Pending
|
|
416
|
+
```
|
|
417
|
+
|
|
418
|
+
### 6.5 ความเข้ากันได้ย้อนหลัง
|
|
419
|
+
|
|
420
|
+
กลไกไฟล์ความคืบหน้ามีความเข้ากันได้ย้อนหลังอย่างสมบูรณ์ —— หากไฟล์ความคืบหน้าไม่มีอยู่ (เช่น โครงการเก่าหรือการทำซ้ำใหม่) Agent ทั้งหมดจะดำเนินการตามตรรกะเดิมตามปกติ
|
|
421
|
+
|
|
422
|
+
---
|
|
423
|
+
|
|
424
|
+
## 7. คำถามที่พบบ่อย (FAQ)
|
|
365
425
|
|
|
366
426
|
### ค1: จะทำอย่างไรหากเอเจนต์ไม่ทำงานตามที่คาดไว้?
|
|
367
427
|
|
|
@@ -389,8 +449,22 @@ iterations/
|
|
|
389
449
|
|
|
390
450
|
### ค4: วิธีอัปเดตเวอร์ชัน SpecCrew?
|
|
391
451
|
|
|
392
|
-
|
|
393
|
-
|
|
452
|
+
การอัปเดตต้องทำสองขั้นตอน:
|
|
453
|
+
|
|
454
|
+
```bash
|
|
455
|
+
# ขั้นตอนที่ 1: อัปเดตเครื่องมือ CLI ระดับโลก
|
|
456
|
+
npm install -g speccrew@latest
|
|
457
|
+
|
|
458
|
+
# ขั้นตอนที่ 2: ซิงค์ Agent และ Skill ในไดเรกทอรีโปรเจกต์ของคุณ
|
|
459
|
+
cd /path/to/your-project
|
|
460
|
+
speccrew update
|
|
461
|
+
```
|
|
462
|
+
|
|
463
|
+
- `npm install -g speccrew@latest`: อัปเดตเครื่องมือ CLI เอง (เวอร์ชันใหม่อาจมีคำจำกัดความ Agent/Skill ใหม่ การแก้ไขบั๊ก ฯลฯ)
|
|
464
|
+
- `speccrew update`: ซิงค์ไฟล์คำจำกัดความ Agent และ Skill ในโปรเจกต์ของคุณเป็นเวอร์ชันล่าสุด
|
|
465
|
+
- `speccrew update --ide cursor`: อัปเดตการกำหนดค่าสำหรับ IDE เฉพาะเท่านั้น
|
|
466
|
+
|
|
467
|
+
> **หมายเหตุ**: จำเป็นต้องทำทั้งสองขั้นตอน การรันเฉพาะ `speccrew update` จะไม่อัปเดตเครื่องมือ CLI เอง การรันเฉพาะ `npm install` จะไม่อัปเดตไฟล์โปรเจกต์
|
|
394
468
|
|
|
395
469
|
### ค5: วิธีดูการทำซ้ำในอดีต?
|
|
396
470
|
|
|
@@ -405,7 +479,7 @@ iterations/
|
|
|
405
479
|
|
|
406
480
|
---
|
|
407
481
|
|
|
408
|
-
##
|
|
482
|
+
## 8. ข้อมูลอ้างอิงด่วน
|
|
409
483
|
|
|
410
484
|
### ข้อมูลอ้างอิงด่วนการเริ่มเอเจนต์
|
|
411
485
|
|
|
@@ -442,7 +516,7 @@ iterations/
|
|
|
442
516
|
|
|
443
517
|
---
|
|
444
518
|
|
|
445
|
-
## ขั้นตอนถัดไป
|
|
519
|
+
## 9. ขั้นตอนถัดไป
|
|
446
520
|
|
|
447
521
|
1. รัน `speccrew init --ide qoder` เพื่อเริ่มต้นโปรเจกต์ของคุณ
|
|
448
522
|
2. ทำขั้นตอนที่ศูนย์: การวินิจฉัยโปรเจกต์และการเริ่มต้นฐานความรู้
|
|
@@ -361,7 +361,67 @@ knowledges/techs/{platform-id}/
|
|
|
361
361
|
|
|
362
362
|
---
|
|
363
363
|
|
|
364
|
-
## 6.
|
|
364
|
+
## 6. İş Akışı İlerleme Yönetimi
|
|
365
|
+
|
|
366
|
+
SpecCrew sanal ekibi, bir sonraki aşamaya geçmeden önce her aşamanın kullanıcı tarafından onaylanması gereken katı bir aşama-geçit mekanizması izler. Ayrıca devam ettirilebilir yürütme desteği sunar — kesinti sonrası yeniden başlatıldığında, kaldığı yerden otomatik olarak devam eder.
|
|
367
|
+
|
|
368
|
+
### 6.1 Üç Katmanlı İlerleme Dosyaları
|
|
369
|
+
|
|
370
|
+
İş akışı, iterasyon dizininde bulunan üç tür JSON ilerleme dosyasını otomatik olarak yönetir:
|
|
371
|
+
|
|
372
|
+
| Dosya | Konum | Amaç |
|
|
373
|
+
|-------|-------|------|
|
|
374
|
+
| `WORKFLOW-PROGRESS.json` | `iterations/{iter}/` | Her boru hattı aşamasının durumunu kaydeder |
|
|
375
|
+
| `.checkpoints.json` | Her aşama dizini altında | Kullanıcı kontrol noktası onay durumunu kaydeder |
|
|
376
|
+
| `DISPATCH-PROGRESS.json` | Her aşama dizini altında | Paralel görevler için öğe bazında ilerlemeyi kaydeder (çoklu platform/çoklu modül) |
|
|
377
|
+
|
|
378
|
+
### 6.2 Aşama Durum Akışı
|
|
379
|
+
|
|
380
|
+
Her aşama bu durum akışını izler:
|
|
381
|
+
|
|
382
|
+
```
|
|
383
|
+
pending → in_progress → completed → confirmed
|
|
384
|
+
```
|
|
385
|
+
|
|
386
|
+
- **pending**: Henüz başlamadı
|
|
387
|
+
- **in_progress**: Şu anda yürütülüyor
|
|
388
|
+
- **completed**: Ajan yürütmesi tamamlandı, kullanıcı onayı bekleniyor
|
|
389
|
+
- **confirmed**: Kullanıcı son kontrol noktası üzerinden onayladı, bir sonraki aşama başlayabilir
|
|
390
|
+
|
|
391
|
+
### 6.3 Devam Ettirilebilir Yürütme
|
|
392
|
+
|
|
393
|
+
Bir aşama için Ajan yeniden başlatıldığında:
|
|
394
|
+
|
|
395
|
+
1. **Otomatik yukarı akış kontrolü**: Önceki aşamanın onaylanıp onaylanmadığını doğrular, değilse engeller ve uyarır
|
|
396
|
+
2. **Kontrol noktası kurtarma**: `.checkpoints.json` dosyasını okur, geçilen kontrol noktalarını atlar, son kesinti noktasından devam eder
|
|
397
|
+
3. **Paralel görev kurtarma**: `DISPATCH-PROGRESS.json` dosyasını okur, yalnızca `pending` veya `failed` durumundaki görevleri yeniden yürütür, `completed` görevleri atlar
|
|
398
|
+
|
|
399
|
+
### 6.4 Mevcut İlerlemeyi Görüntüleme
|
|
400
|
+
|
|
401
|
+
Takım Lideri Ajanı üzerinden boru hattı panoramik durumunu görüntüleyin:
|
|
402
|
+
|
|
403
|
+
```
|
|
404
|
+
@speccrew-team-leader mevcut iterasyon ilerlemesini görüntüle
|
|
405
|
+
```
|
|
406
|
+
|
|
407
|
+
Takım Lideri ilerleme dosyalarını okuyacak ve aşağıdakine benzer bir durum özeti görüntüleyecektir:
|
|
408
|
+
|
|
409
|
+
```
|
|
410
|
+
Pipeline Status: i001-user-management
|
|
411
|
+
01 PRD: ✅ Confirmed
|
|
412
|
+
02 Feature Design: 🔄 In Progress (Checkpoint A passed)
|
|
413
|
+
03 System Design: ⏳ Pending
|
|
414
|
+
04 Development: ⏳ Pending
|
|
415
|
+
05 System Test: ⏳ Pending
|
|
416
|
+
```
|
|
417
|
+
|
|
418
|
+
### 6.5 Geriye Dönük Uyumluluk
|
|
419
|
+
|
|
420
|
+
İlerleme dosyası mekanizması tamamen geriye dönük uyumludur — ilerleme dosyaları mevcut değilse (örneğin, eski projelerde veya yeni iterasyonlarda), tüm Ajanlar orijinal mantığa göre normal şekilde yürütülecektir.
|
|
421
|
+
|
|
422
|
+
---
|
|
423
|
+
|
|
424
|
+
## 7. Sıkça Sorulan Sorular (SSS)
|
|
365
425
|
|
|
366
426
|
### S1: Ajan beklendiği gibi çalışmazsa ne yapmalıyım?
|
|
367
427
|
|
|
@@ -389,8 +449,22 @@ Her iterasyon tamamen izole edilmiştir ve diğerlerini etkilemez.
|
|
|
389
449
|
|
|
390
450
|
### S4: SpecCrew sürümünü nasıl güncellerim?
|
|
391
451
|
|
|
392
|
-
|
|
393
|
-
|
|
452
|
+
Güncelleme iki adım gerektirir:
|
|
453
|
+
|
|
454
|
+
```bash
|
|
455
|
+
# Adım 1: Global CLI aracını güncelle
|
|
456
|
+
npm install -g speccrew@latest
|
|
457
|
+
|
|
458
|
+
# Adım 2: Proje dizininizdeki Ajan ve Becerileri senkronize et
|
|
459
|
+
cd /path/to/your-project
|
|
460
|
+
speccrew update
|
|
461
|
+
```
|
|
462
|
+
|
|
463
|
+
- `npm install -g speccrew@latest`: CLI aracının kendisini günceller (yeni sürümler yeni Ajan/Beceri tanımları, hata düzeltmeleri vb. içerebilir)
|
|
464
|
+
- `speccrew update`: Projenizdeki Ajan ve Beceri tanım dosyalarını en son sürüme senkronize eder
|
|
465
|
+
- `speccrew update --ide cursor`: Yalnızca belirli bir IDE için yapılandırmayı günceller
|
|
466
|
+
|
|
467
|
+
> **Not**: Her iki adım da gereklidir. Yalnızca `speccrew update` çalıştırmak CLI aracının kendisini güncellemez; yalnızca `npm install` çalıştırmak proje dosyalarını güncellemez.
|
|
394
468
|
|
|
395
469
|
### S5: Geçmiş iterasyonları nasıl görüntülerim?
|
|
396
470
|
|
|
@@ -405,7 +479,7 @@ Aşağıdaki durumlarda yeniden başlatma gerekir:
|
|
|
405
479
|
|
|
406
480
|
---
|
|
407
481
|
|
|
408
|
-
##
|
|
482
|
+
## 8. Hızlı Referans
|
|
409
483
|
|
|
410
484
|
### Ajan Başlatma Hızlı Referansı
|
|
411
485
|
|
|
@@ -361,7 +361,67 @@ knowledges/techs/{platform-id}/
|
|
|
361
361
|
|
|
362
362
|
---
|
|
363
363
|
|
|
364
|
-
## 6.
|
|
364
|
+
## 6. Управління Прогресом Робочого Процесу
|
|
365
|
+
|
|
366
|
+
Віртуальна команда SpecCrew дотримується суворого механізму поетапних воріт, де кожен етап повинен бути підтверджений користувачем перед переходом до наступного. Вона також підтримує відновлюване виконання — при перезапуску після переривання автоматично продовжує з того місця, де зупинилася.
|
|
367
|
+
|
|
368
|
+
### 6.1 Трирівневі Файли Прогресу
|
|
369
|
+
|
|
370
|
+
Робочий процес автоматично підтримує три типи JSON файлів прогресу, розташованих в каталозі ітерації:
|
|
371
|
+
|
|
372
|
+
| Файл | Розташування | Призначення |
|
|
373
|
+
|------|--------------|-------------|
|
|
374
|
+
| `WORKFLOW-PROGRESS.json` | `iterations/{iter}/` | Записує статус кожного етапу конвеєра |
|
|
375
|
+
| `.checkpoints.json` | Під кожним каталогом етапу | Записує статус підтвердження контрольної точки користувачем |
|
|
376
|
+
| `DISPATCH-PROGRESS.json` | Під кожним каталогом етапу | Записує поелементний прогрес для паралельних задач (багатоплатформні/багатомодульні) |
|
|
377
|
+
|
|
378
|
+
### 6.2 Потік Статусу Етапу
|
|
379
|
+
|
|
380
|
+
Кожен етап дотримується цього потоку статусу:
|
|
381
|
+
|
|
382
|
+
```
|
|
383
|
+
pending → in_progress → completed → confirmed
|
|
384
|
+
```
|
|
385
|
+
|
|
386
|
+
- **pending**: Ще не розпочато
|
|
387
|
+
- **in_progress**: Наразі виконується
|
|
388
|
+
- **completed**: Виконання Агента завершено, очікує підтвердження користувача
|
|
389
|
+
- **confirmed**: Користувач підтвердив через фінальну контрольну точку, наступний етап може розпочатися
|
|
390
|
+
|
|
391
|
+
### 6.3 Відновлюване Виконання
|
|
392
|
+
|
|
393
|
+
При перезапуску Агента для етапу:
|
|
394
|
+
|
|
395
|
+
1. **Автоматична перевірка вище за течією**: Перевіряє, чи підтверджено попередній етап, блокує та нагадує, якщо ні
|
|
396
|
+
2. **Відновлення контрольної точки**: Читає `.checkpoints.json`, пропускає пройдені контрольні точки, продовжує з останньої точки переривання
|
|
397
|
+
3. **Відновлення паралельних задач**: Читає `DISPATCH-PROGRESS.json`, повторно виконує лише задачі зі статусом `pending` або `failed`, пропускає `completed` задачі
|
|
398
|
+
|
|
399
|
+
### 6.4 Перегляд Поточного Прогресу
|
|
400
|
+
|
|
401
|
+
Перегляд панорамного статусу конвеєра через Агента Лідера Команди:
|
|
402
|
+
|
|
403
|
+
```
|
|
404
|
+
@speccrew-team-leader перегляд прогресу поточної ітерації
|
|
405
|
+
```
|
|
406
|
+
|
|
407
|
+
Лідер Команди прочитає файли прогресу і відобразить огляд статусу, подібний до:
|
|
408
|
+
|
|
409
|
+
```
|
|
410
|
+
Pipeline Status: i001-user-management
|
|
411
|
+
01 PRD: ✅ Confirmed
|
|
412
|
+
02 Feature Design: 🔄 In Progress (Checkpoint A passed)
|
|
413
|
+
03 System Design: ⏳ Pending
|
|
414
|
+
04 Development: ⏳ Pending
|
|
415
|
+
05 System Test: ⏳ Pending
|
|
416
|
+
```
|
|
417
|
+
|
|
418
|
+
### 6.5 Зворотна Сумісність
|
|
419
|
+
|
|
420
|
+
Механізм файлів прогресу повністю зворотно сумісний — якщо файли прогресу не існують (наприклад, в старих проектах або нових ітераціях), всі Агенти виконуватимуться нормально згідно з оригінальною логікою.
|
|
421
|
+
|
|
422
|
+
---
|
|
423
|
+
|
|
424
|
+
## 7. Часті Запитання (FAQ)
|
|
365
425
|
|
|
366
426
|
### П1: Що робити, якщо Агент не працює як очікується?
|
|
367
427
|
|
|
@@ -389,8 +449,22 @@ iterations/
|
|
|
389
449
|
|
|
390
450
|
### П4: Як оновити версію SpecCrew?
|
|
391
451
|
|
|
392
|
-
|
|
393
|
-
|
|
452
|
+
Оновлення вимагає два етапи:
|
|
453
|
+
|
|
454
|
+
```bash
|
|
455
|
+
# Етап 1: Оновити глобальний CLI-інструмент
|
|
456
|
+
npm install -g speccrew@latest
|
|
457
|
+
|
|
458
|
+
# Етап 2: Синхронізувати Агентів та Навички в директорії проекту
|
|
459
|
+
cd /path/to/your-project
|
|
460
|
+
speccrew update
|
|
461
|
+
```
|
|
462
|
+
|
|
463
|
+
- `npm install -g speccrew@latest`: Оновлює сам CLI-інструмент (нові версії можуть включати нові визначення Агентів/Навичок, виправлення помилок тощо)
|
|
464
|
+
- `speccrew update`: Синхронізує файли визначень Агентів та Навичок у вашому проекті до останньої версії
|
|
465
|
+
- `speccrew update --ide cursor`: Оновлює конфігурацію тільки для конкретної IDE
|
|
466
|
+
|
|
467
|
+
> **Примітка**: Обидва етапи обов'язкові. Виконання тільки `speccrew update` не оновить сам CLI-інструмент; виконання тільки `npm install` не оновить файли проекту.
|
|
394
468
|
|
|
395
469
|
### П5: Як переглянути історичні ітерації?
|
|
396
470
|
|
|
@@ -405,7 +479,7 @@ iterations/
|
|
|
405
479
|
|
|
406
480
|
---
|
|
407
481
|
|
|
408
|
-
##
|
|
482
|
+
## 8. Швидка Довідка
|
|
409
483
|
|
|
410
484
|
### Швидка Довідка Запуску Агентів
|
|
411
485
|
|
|
@@ -361,7 +361,67 @@ knowledges/techs/{platform-id}/
|
|
|
361
361
|
|
|
362
362
|
---
|
|
363
363
|
|
|
364
|
-
## 6.
|
|
364
|
+
## 6. Quản Lý Tiến Độ Pipeline
|
|
365
|
+
|
|
366
|
+
Nhóm ảo SpecCrew tuân thủ cơ chế kiểm soát pha nghiêm ngặt, mỗi pha phải được xác nhận bởi ngườ dùng trước khi có thể tiến hành sang pha tiếp theo. Đồng thờ hỗ trợ tiếp tục từ điểm dừng —— khi khởi động lại sau khi bị gián đoạn, hệ thống sẽ tự động tiếp tục từ vị trí dừng lần cuối.
|
|
367
|
+
|
|
368
|
+
### 6.1 Ba Tệp Tiến Độ
|
|
369
|
+
|
|
370
|
+
Workflow tự động duy trì ba loại tệp tiến độ JSON, nằm trong thư mục iteration:
|
|
371
|
+
|
|
372
|
+
| Tệp | Vị trí | Vai trò |
|
|
373
|
+
|------|----------|--------|
|
|
374
|
+
| `WORKFLOW-PROGRESS.json` | `iterations/{iter}/` | Ghi lại trạng thái của từng pha trong toàn bộ pipeline |
|
|
375
|
+
| `.checkpoints.json` | Trong thư mục từng pha | Ghi lại trạng thái vượt qua điểm kiểm tra (Checkpoint) của ngườ dùng |
|
|
376
|
+
| `DISPATCH-PROGRESS.json` | Trong thư mục từng pha | Ghi lại tiến độ từng mục của tác vụ song song (đa nền tảng/module) |
|
|
377
|
+
|
|
378
|
+
### 6.2 Chuyển Đổi Trạng Thái Pha
|
|
379
|
+
|
|
380
|
+
Mỗi pha tuân thủ chuyển đổi trạng thái sau:
|
|
381
|
+
|
|
382
|
+
```
|
|
383
|
+
pending → in_progress → completed → confirmed
|
|
384
|
+
```
|
|
385
|
+
|
|
386
|
+
- **pending**: Chưa bắt đầu
|
|
387
|
+
- **in_progress**: Đang thực hiện
|
|
388
|
+
- **completed**: Agent thực hiện xong, chờ xác nhận từ ngườ dùng
|
|
389
|
+
- **confirmed**: Ngườ dùng xác nhận Checkpoint cuối cùng, có thể bắt đầu pha tiếp theo
|
|
390
|
+
|
|
391
|
+
### 6.3 Tiếp Tục Từ Điểm Dừng
|
|
392
|
+
|
|
393
|
+
Khi khởi động lại Agent của một pha nào đó:
|
|
394
|
+
|
|
395
|
+
1. **Tự động kiểm tra upstream**: Xác minh pha trước đã được confirmed chưa, nếu chưa sẽ bị chặn và cảnh báo
|
|
396
|
+
2. **Khôi phục Checkpoint**: Đọc `.checkpoints.json`, bỏ qua các điểm kiểm tra đã vượt qua, và tiếp tục từ điểm dừng lần cuối
|
|
397
|
+
3. **Khôi phục tác vụ song song**: Đọc `DISPATCH-PROGRESS.json`, chỉ thực thi lại các tác vụ có trạng thái `pending` hoặc `failed`, bỏ qua các tác vụ đã `completed`
|
|
398
|
+
|
|
399
|
+
### 6.4 Xem Tiến Độ Hiện Tại
|
|
400
|
+
|
|
401
|
+
Xem trạng thái toàn cảnh của pipeline thông qua Team Leader Agent:
|
|
402
|
+
|
|
403
|
+
```
|
|
404
|
+
@speccrew-team-leader xem tiến độ iteration hiện tại
|
|
405
|
+
```
|
|
406
|
+
|
|
407
|
+
Team Leader sẽ đọc các tệp tiến độ và hiển thị tóm tắt trạng thái tương tự như sau:
|
|
408
|
+
|
|
409
|
+
```
|
|
410
|
+
Pipeline Status: i001-user-management
|
|
411
|
+
01 PRD: ✅ Confirmed
|
|
412
|
+
02 Feature Design: 🔄 In Progress (Checkpoint A passed)
|
|
413
|
+
03 System Design: ⏳ Pending
|
|
414
|
+
04 Development: ⏳ Pending
|
|
415
|
+
05 System Test: ⏳ Pending
|
|
416
|
+
```
|
|
417
|
+
|
|
418
|
+
### 6.5 Tương Thích Ngược
|
|
419
|
+
|
|
420
|
+
Cơ chế tệp tiến độ hoàn toàn tương thích ngược —— nếu tệp tiến độ không tồn tại (ví dụ: dự án cũ hoặc iteration mới), tất cả các Agent sẽ thực thi theo logic ban đầu như bình thường.
|
|
421
|
+
|
|
422
|
+
---
|
|
423
|
+
|
|
424
|
+
## 7. Câu Hỏi Thường Gặp (FAQ)
|
|
365
425
|
|
|
366
426
|
### H1: Phải làm gì nếu Agent không hoạt động như mong đợi?
|
|
367
427
|
|
|
@@ -389,8 +449,22 @@ Mỗi iteration được cô lập hoàn toàn và không ảnh hưởng đến
|
|
|
389
449
|
|
|
390
450
|
### H4: Làm thế nào để cập nhật phiên bản SpecCrew?
|
|
391
451
|
|
|
392
|
-
|
|
393
|
-
|
|
452
|
+
Cập nhật yêu cầu hai bước:
|
|
453
|
+
|
|
454
|
+
```bash
|
|
455
|
+
# Bước 1: Cập nhật công cụ CLI toàn cục
|
|
456
|
+
npm install -g speccrew@latest
|
|
457
|
+
|
|
458
|
+
# Bước 2: Đồng bộ Agent và Skill trong thư mục dự án của bạn
|
|
459
|
+
cd /path/to/your-project
|
|
460
|
+
speccrew update
|
|
461
|
+
```
|
|
462
|
+
|
|
463
|
+
- `npm install -g speccrew@latest`: Cập nhật chính công cụ CLI (phiên bản mới có thể bao gồm định nghĩa Agent/Skill mới, sửa lỗi, v.v.)
|
|
464
|
+
- `speccrew update`: Đồng bộ các tệp định nghĩa Agent và Skill trong dự án của bạn lên phiên bản mới nhất
|
|
465
|
+
- `speccrew update --ide cursor`: Cập nhật cấu hình chỉ cho một IDE cụ thể
|
|
466
|
+
|
|
467
|
+
> **Lưu ý**: Cả hai bước đều cần thiết. Chỉ chạy `speccrew update` sẽ không cập nhật chính công cụ CLI; chỉ chạy `npm install` sẽ không cập nhật các tệp dự án.
|
|
394
468
|
|
|
395
469
|
### H5: Làm thế nào để xem các iteration lịch sử?
|
|
396
470
|
|
|
@@ -405,7 +479,7 @@ Cần khởi tạo lại trong các tình huống sau:
|
|
|
405
479
|
|
|
406
480
|
---
|
|
407
481
|
|
|
408
|
-
##
|
|
482
|
+
## 8. Tham Khảo Nhanh
|
|
409
483
|
|
|
410
484
|
### Tham Khảo Nhanh Khởi Động Agent
|
|
411
485
|
|
|
@@ -442,7 +516,7 @@ Cần khởi tạo lại trong các tình huống sau:
|
|
|
442
516
|
|
|
443
517
|
---
|
|
444
518
|
|
|
445
|
-
## Bước Tiếp Theo
|
|
519
|
+
## 9. Bước Tiếp Theo
|
|
446
520
|
|
|
447
521
|
1. Chạy `speccrew init --ide qoder` để khởi tạo dự án của bạn
|
|
448
522
|
2. Thực hiện Bước Không: Chẩn Đoán Dự Án và Khởi Tạo Cơ Sở Kiến Thức
|
|
@@ -360,7 +360,67 @@ knowledges/techs/{platform-id}/
|
|
|
360
360
|
|
|
361
361
|
---
|
|
362
362
|
|
|
363
|
-
## 6.
|
|
363
|
+
## 6. 流水線進度管理
|
|
364
|
+
|
|
365
|
+
SpecCrew 虛擬團隊遵循嚴格的階段門控機制,每個階段必須經過用戶確認後才能推進到下一階段。同時支援斷點續傳 —— 中斷後重新啟動時,自動從上次停止的位置繼續。
|
|
366
|
+
|
|
367
|
+
### 6.1 三層進度文件
|
|
368
|
+
|
|
369
|
+
工作流自動維護三類 JSON 進度文件,位於迭代目錄下:
|
|
370
|
+
|
|
371
|
+
| 文件 | 位置 | 作用 |
|
|
372
|
+
|------|------|------|
|
|
373
|
+
| `WORKFLOW-PROGRESS.json` | `iterations/{iter}/` | 記錄整條流水線各階段狀態 |
|
|
374
|
+
| `.checkpoints.json` | 各階段目錄下 | 記錄用戶確認點(Checkpoint)通過狀態 |
|
|
375
|
+
| `DISPATCH-PROGRESS.json` | 各階段目錄下 | 記錄並行任務(多平台/多模組)的逐項進度 |
|
|
376
|
+
|
|
377
|
+
### 6.2 階段狀態流轉
|
|
378
|
+
|
|
379
|
+
每個階段遵循以下狀態流轉:
|
|
380
|
+
|
|
381
|
+
```
|
|
382
|
+
pending → in_progress → completed → confirmed
|
|
383
|
+
```
|
|
384
|
+
|
|
385
|
+
- **pending**:尚未開始
|
|
386
|
+
- **in_progress**:正在執行中
|
|
387
|
+
- **completed**:Agent 執行完成,等待用戶確認
|
|
388
|
+
- **confirmed**:用戶通過最終 Checkpoint 確認,下一階段可以啟動
|
|
389
|
+
|
|
390
|
+
### 6.3 斷點續傳
|
|
391
|
+
|
|
392
|
+
當重新啟動某個階段的 Agent 時:
|
|
393
|
+
|
|
394
|
+
1. **自動檢查上游**:驗證前一階段是否已 confirmed,未確認則阻塞並提示
|
|
395
|
+
2. **恢復 Checkpoint**:讀取 `.checkpoints.json`,跳過已通過的確認點,從上次中斷處繼續
|
|
396
|
+
3. **恢復並行任務**:讀取 `DISPATCH-PROGRESS.json`,只重新執行 `pending` 或 `failed` 狀態的任務,跳過已 `completed` 的任務
|
|
397
|
+
|
|
398
|
+
### 6.4 查看當前進度
|
|
399
|
+
|
|
400
|
+
通過 Team Leader Agent 查看流水線全景狀態:
|
|
401
|
+
|
|
402
|
+
```
|
|
403
|
+
@speccrew-team-leader 查看當前迭代進度
|
|
404
|
+
```
|
|
405
|
+
|
|
406
|
+
Team Leader 會讀取進度文件並展示類似以下的狀態概覽:
|
|
407
|
+
|
|
408
|
+
```
|
|
409
|
+
Pipeline Status: i001-user-management
|
|
410
|
+
01 PRD: ✅ Confirmed
|
|
411
|
+
02 Feature Design: 🔄 In Progress (Checkpoint A passed)
|
|
412
|
+
03 System Design: ⏳ Pending
|
|
413
|
+
04 Development: ⏳ Pending
|
|
414
|
+
05 System Test: ⏳ Pending
|
|
415
|
+
```
|
|
416
|
+
|
|
417
|
+
### 6.5 向下兼容
|
|
418
|
+
|
|
419
|
+
進度文件機制完全向下兼容 —— 如果進度文件不存在(如舊項目或全新迭代),所有 Agent 將按照原有邏輯正常執行。
|
|
420
|
+
|
|
421
|
+
---
|
|
422
|
+
|
|
423
|
+
## 7. 常見問題(FAQ)
|
|
364
424
|
|
|
365
425
|
### Q1: Agent 不按預期工作怎麼辦?
|
|
366
426
|
|
|
@@ -388,8 +448,22 @@ iterations/
|
|
|
388
448
|
|
|
389
449
|
### Q4: 如何更新 SpecCrew 版本?
|
|
390
450
|
|
|
391
|
-
|
|
392
|
-
|
|
451
|
+
更新需要兩個步驟:
|
|
452
|
+
|
|
453
|
+
```bash
|
|
454
|
+
# 步驟一:更新全域 CLI 工具
|
|
455
|
+
npm install -g speccrew@latest
|
|
456
|
+
|
|
457
|
+
# 步驟二:在專案目錄同步 Agent 和 Skill
|
|
458
|
+
cd /path/to/your-project
|
|
459
|
+
speccrew update
|
|
460
|
+
```
|
|
461
|
+
|
|
462
|
+
- `npm install -g speccrew@latest`:更新 CLI 工具本身(新版本可能包含新的 Agent/Skill 定義、錯誤修復等)
|
|
463
|
+
- `speccrew update`:將專案中的 Agent 和 Skill 定義檔案同步到最新版本
|
|
464
|
+
- `speccrew update --ide cursor`:僅更新特定 IDE 的設定
|
|
465
|
+
|
|
466
|
+
> **注意**:兩個步驟都需要執行。只執行 `speccrew update` 不會更新 CLI 工具本身;只執行 `npm install` 不會更新專案檔案。
|
|
393
467
|
|
|
394
468
|
### Q5: 如何查看歷史迭代?
|
|
395
469
|
|
|
@@ -404,7 +478,7 @@ iterations/
|
|
|
404
478
|
|
|
405
479
|
---
|
|
406
480
|
|
|
407
|
-
##
|
|
481
|
+
## 8. 快速參考
|
|
408
482
|
|
|
409
483
|
### Agent 啟動速查表
|
|
410
484
|
|
|
@@ -441,7 +515,7 @@ iterations/
|
|
|
441
515
|
|
|
442
516
|
---
|
|
443
517
|
|
|
444
|
-
## 下一步
|
|
518
|
+
## 9. 下一步
|
|
445
519
|
|
|
446
520
|
1. 執行 `speccrew init --ide qoder` 初始化您的專案
|
|
447
521
|
2. 執行第零步:專案診斷與知識庫初始化
|
package/package.json
CHANGED
|
@@ -1,98 +0,0 @@
|
|
|
1
|
-
---
|
|
2
|
-
name: speccrew-create-agents
|
|
3
|
-
description: Creates or updates tech-stack-specific Agents and project-level Skills based on project diagnosis report. Generates role Agents using predefined templates. Use after project diagnosis completion when setting up or updating the AI collaboration system.
|
|
4
|
-
---
|
|
5
|
-
|
|
6
|
-
# Create Agents and Skills
|
|
7
|
-
|
|
8
|
-
Based on the diagnosis report in `speccrew-workspace/diagnosis-reports/`, generate or update tech-stack-specific Agents and project-level Skills for the project.
|
|
9
|
-
|
|
10
|
-
## Prerequisites
|
|
11
|
-
|
|
12
|
-
**Must complete project diagnosis first**, ensure `speccrew-workspace/diagnosis-reports/diagnosis-report-{date}.md` exists and contains complete information. If not exists, prompt user to execute `speccrew-project-diagnosis` Skill first.
|
|
13
|
-
|
|
14
|
-
**Read the latest diagnosis report**:
|
|
15
|
-
1. List all diagnosis report files in `speccrew-workspace/diagnosis-reports/` directory
|
|
16
|
-
2. Sort by filename date, select the latest report
|
|
17
|
-
3. Read report content as generation basis
|
|
18
|
-
|
|
19
|
-
## Attached Resources
|
|
20
|
-
|
|
21
|
-
This Skill directory contains the following predefined files:
|
|
22
|
-
|
|
23
|
-
**Agent Templates** (`templates/agents/`, used for creating tech-stack-specific Agents):
|
|
24
|
-
- [templates/agents/designer-agent.md](templates/agents/designer-agent.md): Detailed Design Agent template
|
|
25
|
-
- [templates/agents/dev-agent.md](templates/agents/dev-agent.md): Development Agent template
|
|
26
|
-
- [templates/agents/test-agent.md](templates/agents/test-agent.md): Testing Agent template
|
|
27
|
-
|
|
28
|
-
**Note**: Use these templates to create Agents by replacing `[techstack]` placeholder with actual technology stack from diagnosis report.
|
|
29
|
-
|
|
30
|
-
## Preconditions
|
|
31
|
-
|
|
32
|
-
Before execution, confirm the following has been obtained from the latest diagnosis report:
|
|
33
|
-
- **Project Type** (Web Full-Stack / Frontend Only / Backend Only / Desktop Client / Mobile / CLI / Hybrid)
|
|
34
|
-
- Actual technology stack used (language versions, core frameworks, database)
|
|
35
|
-
- Directory conventions (actual paths where various files are stored)
|
|
36
|
-
- Code standards (lint tools, naming conventions, run commands)
|
|
37
|
-
|
|
38
|
-
## Execution Steps
|
|
39
|
-
|
|
40
|
-
### Step 1: Check Existing Files
|
|
41
|
-
|
|
42
|
-
Scan `agents/` and `skills/` directories, record existing files. For existing Agent files, they will be updated (not overwritten) based on latest diagnosis report in subsequent steps.
|
|
43
|
-
|
|
44
|
-
### Step 2: Generate Tech-Stack-Specific Agents
|
|
45
|
-
|
|
46
|
-
Read the **Recommended Agents to Generate** section from diagnosis report, create or update tech-stack-specific Agents in `agents/`:
|
|
47
|
-
|
|
48
|
-
- `speccrew-designer-[techstack]` (e.g., speccrew-designer-react, speccrew-designer-fastapi) - use `templates/agents/designer-agent.md`
|
|
49
|
-
- `speccrew-dev-[techstack]` (e.g., speccrew-dev-nextjs, speccrew-dev-springboot) - use `templates/agents/dev-agent.md`
|
|
50
|
-
- `speccrew-test-[techstack]` (e.g., speccrew-test-playwright, speccrew-test-junit) - use `templates/agents/test-agent.md`
|
|
51
|
-
|
|
52
|
-
**Template Usage**:
|
|
53
|
-
- **If Agent does not exist**: Copy template, replace `[techstack]` placeholder with actual technology stack name from diagnosis report, embed project-specific information
|
|
54
|
-
- **If Agent already exists**: Read existing Agent file, update technology stack info, directory paths, commands, and standards based on latest diagnosis report (preserve existing workflow logic)
|
|
55
|
-
|
|
56
|
-
**Note**: Generic agents (leader-agent,pm-agent, solution-agent) are created during project initialization, not here.
|
|
57
|
-
|
|
58
|
-
**Each Agent file must embed project actual information:**
|
|
59
|
-
- Actual technology stack name and version (from diagnosis report)
|
|
60
|
-
- Actual directory paths (from diagnosis report, no guessing)
|
|
61
|
-
- Actual run/debug commands (from diagnosis report)
|
|
62
|
-
- Actual code standard requirements (from diagnosis report)
|
|
63
|
-
|
|
64
|
-
### Step 3: Generate Project-Level Skill Files
|
|
65
|
-
|
|
66
|
-
Create or update project-level Skill directory structure in `skills/`.
|
|
67
|
-
|
|
68
|
-
**Note**: Specific Skill content (such as add-page, add-api, etc.) is not generated at this stage, but gradually created by Dev Agent through `speccrew-skill-develop` Skill after identifying repetitive operation patterns during development.
|
|
69
|
-
|
|
70
|
-
This step only creates necessary directory structure and basic configuration Skills (such as pre-commit-check, if configured). Existing Skill files are preserved.
|
|
71
|
-
|
|
72
|
-
### Step 4: Output Generation Summary
|
|
73
|
-
|
|
74
|
-
List:
|
|
75
|
-
- Created files list (path + description)
|
|
76
|
-
- Skipped files list (path + reason)
|
|
77
|
-
- Suggested next steps
|
|
78
|
-
|
|
79
|
-
## Agent File Content Standards
|
|
80
|
-
|
|
81
|
-
Each Agent file must:
|
|
82
|
-
1. **Embed project-specific information**, no generalized descriptions
|
|
83
|
-
2. Clarify "context input" (which directory and what type of files to read)
|
|
84
|
-
3. Clarify "output standards" (where to store deliverables)
|
|
85
|
-
4. Include "escalate when ambiguity found" constraint
|
|
86
|
-
|
|
87
|
-
## Skill File Content Standards
|
|
88
|
-
|
|
89
|
-
Each Skill file must:
|
|
90
|
-
1. Contain **specific file paths** (e.g., `server/routers/` instead of "router directory")
|
|
91
|
-
2. Contain **verification checklist** (how to confirm operation success after completion)
|
|
92
|
-
3. Keep step count within 10 steps
|
|
93
|
-
|
|
94
|
-
## Notes
|
|
95
|
-
|
|
96
|
-
- Check if files already exist before writing, skip if exists (no overwrite)
|
|
97
|
-
- Agent description field must contain clear "when to trigger" explanation
|
|
98
|
-
|