takt 0.1.3 → 0.1.5
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/dist/agents/runner.d.ts +2 -0
- package/dist/agents/runner.d.ts.map +1 -1
- package/dist/agents/runner.js +10 -2
- package/dist/agents/runner.js.map +1 -1
- package/dist/claude/client.d.ts.map +1 -1
- package/dist/claude/client.js +4 -1
- package/dist/claude/client.js.map +1 -1
- package/dist/claude/executor.d.ts.map +1 -1
- package/dist/claude/executor.js +1 -0
- package/dist/claude/executor.js.map +1 -1
- package/dist/claude/types.d.ts +2 -0
- package/dist/claude/types.d.ts.map +1 -1
- package/dist/commands/workflowExecution.d.ts.map +1 -1
- package/dist/commands/workflowExecution.js +25 -2
- package/dist/commands/workflowExecution.js.map +1 -1
- package/dist/config/workflowLoader.d.ts.map +1 -1
- package/dist/config/workflowLoader.js +1 -0
- package/dist/config/workflowLoader.js.map +1 -1
- package/dist/models/schemas.d.ts +2 -0
- package/dist/models/schemas.d.ts.map +1 -1
- package/dist/models/schemas.js +1 -0
- package/dist/models/schemas.js.map +1 -1
- package/dist/models/types.d.ts +4 -0
- package/dist/models/types.d.ts.map +1 -1
- package/dist/workflow/engine.d.ts.map +1 -1
- package/dist/workflow/engine.js +7 -3
- package/dist/workflow/engine.js.map +1 -1
- package/dist/workflow/instruction-builder.d.ts +5 -2
- package/dist/workflow/instruction-builder.d.ts.map +1 -1
- package/dist/workflow/instruction-builder.js +8 -2
- package/dist/workflow/instruction-builder.js.map +1 -1
- package/dist/workflow/state-manager.d.ts +4 -0
- package/dist/workflow/state-manager.d.ts.map +1 -1
- package/dist/workflow/state-manager.js +10 -0
- package/dist/workflow/state-manager.js.map +1 -1
- package/package.json +1 -1
- package/resources/global/en/agents/default/ai-reviewer.md +0 -20
- package/resources/global/en/agents/default/architect.md +5 -73
- package/resources/global/en/agents/default/coder.md +0 -39
- package/resources/global/en/agents/default/planner.md +0 -23
- package/resources/global/en/agents/default/security.md +0 -16
- package/resources/global/en/agents/default/supervisor.md +0 -19
- package/resources/global/en/agents/expert-review/cqrs-es-reviewer.md +0 -35
- package/resources/global/en/agents/expert-review/frontend-reviewer.md +0 -35
- package/resources/global/en/agents/expert-review/qa-reviewer.md +0 -36
- package/resources/global/en/agents/expert-review/security-reviewer.md +0 -37
- package/resources/global/en/agents/expert-review/supervisor.md +0 -62
- package/resources/global/en/agents/magi/balthasar.md +0 -20
- package/resources/global/en/agents/magi/casper.md +0 -42
- package/resources/global/en/agents/magi/melchior.md +0 -20
- package/resources/global/en/agents/research/digger.md +0 -41
- package/resources/global/en/agents/research/planner.md +0 -34
- package/resources/global/en/agents/research/supervisor.md +0 -36
- package/resources/global/en/workflows/default.yaml +419 -46
- package/resources/global/en/workflows/expert-review.yaml +217 -22
- package/resources/global/en/workflows/magi.yaml +78 -26
- package/resources/global/en/workflows/research.yaml +105 -11
- package/resources/global/ja/agents/default/ai-reviewer.md +0 -20
- package/resources/global/ja/agents/default/architect.md +5 -73
- package/resources/global/ja/agents/default/coder.md +1 -41
- package/resources/global/ja/agents/default/planner.md +0 -23
- package/resources/global/ja/agents/default/security.md +0 -16
- package/resources/global/ja/agents/default/supervisor.md +0 -19
- package/resources/global/ja/agents/expert-review/cqrs-es-reviewer.md +0 -35
- package/resources/global/ja/agents/expert-review/frontend-reviewer.md +0 -35
- package/resources/global/ja/agents/expert-review/qa-reviewer.md +0 -36
- package/resources/global/ja/agents/expert-review/security-reviewer.md +0 -37
- package/resources/global/ja/agents/expert-review/supervisor.md +0 -62
- package/resources/global/ja/agents/magi/balthasar.md +0 -20
- package/resources/global/ja/agents/magi/casper.md +0 -42
- package/resources/global/ja/agents/magi/melchior.md +0 -20
- package/resources/global/ja/agents/research/digger.md +0 -41
- package/resources/global/ja/agents/research/planner.md +0 -34
- package/resources/global/ja/agents/research/supervisor.md +0 -36
- package/resources/global/ja/workflows/default.yaml +419 -50
- package/resources/global/ja/workflows/expert-review.yaml +217 -22
- package/resources/global/ja/workflows/magi.yaml +78 -26
- package/resources/global/ja/workflows/research.yaml +105 -11
|
@@ -9,6 +9,16 @@
|
|
|
9
9
|
# Fix destination is determined by Coder based on change impact:
|
|
10
10
|
# - fix_security: MINOR→security_review, MAJOR→cqrs_es_review
|
|
11
11
|
# - fix_qa: MINOR→qa_review, SECURITY→security_review, MAJOR→cqrs_es_review
|
|
12
|
+
#
|
|
13
|
+
# Template Variables:
|
|
14
|
+
# {iteration} - Workflow-wide turn count (total steps executed across all agents)
|
|
15
|
+
# {max_iterations} - Maximum iterations allowed for the workflow
|
|
16
|
+
# {step_iteration} - Per-step iteration count (how many times THIS step has been executed)
|
|
17
|
+
# {task} - Original user request
|
|
18
|
+
# {previous_response} - Output from the previous step
|
|
19
|
+
# {git_diff} - Current uncommitted changes (git diff)
|
|
20
|
+
# {user_inputs} - Accumulated user inputs during workflow
|
|
21
|
+
# {report_dir} - Report directory name (e.g., "20250126-143052-task-summary")
|
|
12
22
|
|
|
13
23
|
name: expert-review
|
|
14
24
|
description: CQRS+ES, Frontend, Security, QA Expert Review
|
|
@@ -30,9 +40,22 @@ steps:
|
|
|
30
40
|
- Bash
|
|
31
41
|
- WebSearch
|
|
32
42
|
- WebFetch
|
|
43
|
+
status_rules_prompt: |
|
|
44
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
45
|
+
|
|
46
|
+
**Without this tag, the workflow will stop.**
|
|
47
|
+
Your final output MUST include a status tag following the rules below.
|
|
48
|
+
|
|
49
|
+
## Output Format
|
|
50
|
+
|
|
51
|
+
| Situation | Tag |
|
|
52
|
+
|-----------|-----|
|
|
53
|
+
| Analysis complete | `[PLANNER:DONE]` |
|
|
54
|
+
| Requirements unclear | `[PLANNER:BLOCKED]` |
|
|
33
55
|
instruction_template: |
|
|
34
56
|
## Workflow Context
|
|
35
|
-
- Iteration: {iteration}/{max_iterations}
|
|
57
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
58
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
36
59
|
- Step: plan (Task Analysis)
|
|
37
60
|
- Report Directory: .takt/reports/{report_dir}/
|
|
38
61
|
- Report File: .takt/reports/{report_dir}/00-plan.md
|
|
@@ -56,7 +79,7 @@ steps:
|
|
|
56
79
|
|
|
57
80
|
**Report output:** Output to the `Report File` specified above.
|
|
58
81
|
- If file does not exist: Create new file
|
|
59
|
-
- If file exists: Append with `## Iteration {
|
|
82
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
60
83
|
|
|
61
84
|
**Report format:**
|
|
62
85
|
```markdown
|
|
@@ -103,9 +126,22 @@ steps:
|
|
|
103
126
|
- Bash
|
|
104
127
|
- WebSearch
|
|
105
128
|
- WebFetch
|
|
129
|
+
status_rules_prompt: |
|
|
130
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
131
|
+
|
|
132
|
+
**Without this tag, the workflow will stop.**
|
|
133
|
+
Your final output MUST include a status tag following the rules below.
|
|
134
|
+
|
|
135
|
+
## Output Format
|
|
136
|
+
|
|
137
|
+
| Situation | Tag |
|
|
138
|
+
|-----------|-----|
|
|
139
|
+
| Implementation complete | `[CODER:DONE]` |
|
|
140
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
106
141
|
instruction_template: |
|
|
107
142
|
## Workflow Context
|
|
108
|
-
- Iteration: {iteration}/{max_iterations}
|
|
143
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
144
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
109
145
|
- Step: implement
|
|
110
146
|
- Report Directory: .takt/reports/{report_dir}/
|
|
111
147
|
- Report Files:
|
|
@@ -124,7 +160,7 @@ steps:
|
|
|
124
160
|
|
|
125
161
|
**Report output:** Output to the `Report Files` specified above.
|
|
126
162
|
- If file does not exist: Create new file
|
|
127
|
-
- If file exists: Append with `## Iteration {
|
|
163
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
128
164
|
|
|
129
165
|
**Scope report format (create at implementation start):**
|
|
130
166
|
```markdown
|
|
@@ -175,9 +211,22 @@ steps:
|
|
|
175
211
|
- Grep
|
|
176
212
|
- WebSearch
|
|
177
213
|
- WebFetch
|
|
214
|
+
status_rules_prompt: |
|
|
215
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
216
|
+
|
|
217
|
+
**Without this tag, the workflow will stop.**
|
|
218
|
+
Your final output MUST include a status tag following the rules below.
|
|
219
|
+
|
|
220
|
+
## Output Format
|
|
221
|
+
|
|
222
|
+
| Situation | Tag |
|
|
223
|
+
|-----------|-----|
|
|
224
|
+
| Design is sound | `[CQRS-ES:APPROVE]` |
|
|
225
|
+
| Design issues found | `[CQRS-ES:REJECT]` |
|
|
178
226
|
instruction_template: |
|
|
179
227
|
## Workflow Context
|
|
180
|
-
- Iteration: {iteration}/{max_iterations}
|
|
228
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
229
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
181
230
|
- Step: cqrs_es_review (CQRS+ES Expert Review)
|
|
182
231
|
- Report Directory: .takt/reports/{report_dir}/
|
|
183
232
|
- Report File: .takt/reports/{report_dir}/03-cqrs-es-review.md
|
|
@@ -206,7 +255,7 @@ steps:
|
|
|
206
255
|
|
|
207
256
|
**Report output:** Output to the `Report File` specified above.
|
|
208
257
|
- If file does not exist: Create new file
|
|
209
|
-
- If file exists: Append with `## Iteration {
|
|
258
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
210
259
|
|
|
211
260
|
**Report format:**
|
|
212
261
|
```markdown
|
|
@@ -252,9 +301,22 @@ steps:
|
|
|
252
301
|
- Bash
|
|
253
302
|
- WebSearch
|
|
254
303
|
- WebFetch
|
|
304
|
+
status_rules_prompt: |
|
|
305
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
306
|
+
|
|
307
|
+
**Without this tag, the workflow will stop.**
|
|
308
|
+
Your final output MUST include a status tag following the rules below.
|
|
309
|
+
|
|
310
|
+
## Output Format
|
|
311
|
+
|
|
312
|
+
| Situation | Tag |
|
|
313
|
+
|-----------|-----|
|
|
314
|
+
| Fix complete | `[CODER:DONE]` |
|
|
315
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
255
316
|
instruction_template: |
|
|
256
317
|
## Workflow Context
|
|
257
|
-
- Iteration: {iteration}/{max_iterations}
|
|
318
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
319
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
258
320
|
- Step: fix_cqrs_es
|
|
259
321
|
|
|
260
322
|
## CQRS+ES Review Feedback (This is the latest instruction - prioritize this)
|
|
@@ -296,9 +358,22 @@ steps:
|
|
|
296
358
|
- Grep
|
|
297
359
|
- WebSearch
|
|
298
360
|
- WebFetch
|
|
361
|
+
status_rules_prompt: |
|
|
362
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
363
|
+
|
|
364
|
+
**Without this tag, the workflow will stop.**
|
|
365
|
+
Your final output MUST include a status tag following the rules below.
|
|
366
|
+
|
|
367
|
+
## Output Format
|
|
368
|
+
|
|
369
|
+
| Situation | Tag |
|
|
370
|
+
|-----------|-----|
|
|
371
|
+
| Frontend design is sound | `[FRONTEND:APPROVE]` |
|
|
372
|
+
| Design issues found | `[FRONTEND:REJECT]` |
|
|
299
373
|
instruction_template: |
|
|
300
374
|
## Workflow Context
|
|
301
|
-
- Iteration: {iteration}/{max_iterations}
|
|
375
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
376
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
302
377
|
- Step: frontend_review (Frontend Expert Review)
|
|
303
378
|
- Report Directory: .takt/reports/{report_dir}/
|
|
304
379
|
- Report File: .takt/reports/{report_dir}/04-frontend-review.md
|
|
@@ -327,7 +402,7 @@ steps:
|
|
|
327
402
|
|
|
328
403
|
**Report output:** Output to the `Report File` specified above.
|
|
329
404
|
- If file does not exist: Create new file
|
|
330
|
-
- If file exists: Append with `## Iteration {
|
|
405
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
331
406
|
|
|
332
407
|
**Report format:**
|
|
333
408
|
```markdown
|
|
@@ -373,9 +448,22 @@ steps:
|
|
|
373
448
|
- Bash
|
|
374
449
|
- WebSearch
|
|
375
450
|
- WebFetch
|
|
451
|
+
status_rules_prompt: |
|
|
452
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
453
|
+
|
|
454
|
+
**Without this tag, the workflow will stop.**
|
|
455
|
+
Your final output MUST include a status tag following the rules below.
|
|
456
|
+
|
|
457
|
+
## Output Format
|
|
458
|
+
|
|
459
|
+
| Situation | Tag |
|
|
460
|
+
|-----------|-----|
|
|
461
|
+
| Fix complete | `[CODER:DONE]` |
|
|
462
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
376
463
|
instruction_template: |
|
|
377
464
|
## Workflow Context
|
|
378
|
-
- Iteration: {iteration}/{max_iterations}
|
|
465
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
466
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
379
467
|
- Step: fix_frontend
|
|
380
468
|
|
|
381
469
|
## Frontend Review Feedback (This is the latest instruction - prioritize this)
|
|
@@ -417,9 +505,22 @@ steps:
|
|
|
417
505
|
- Grep
|
|
418
506
|
- WebSearch
|
|
419
507
|
- WebFetch
|
|
508
|
+
status_rules_prompt: |
|
|
509
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
510
|
+
|
|
511
|
+
**Without this tag, the workflow will stop.**
|
|
512
|
+
Your final output MUST include a status tag following the rules below.
|
|
513
|
+
|
|
514
|
+
## Output Format
|
|
515
|
+
|
|
516
|
+
| Situation | Tag |
|
|
517
|
+
|-----------|-----|
|
|
518
|
+
| No AI-specific issues | `[AI_REVIEW:APPROVE]` |
|
|
519
|
+
| Issues found | `[AI_REVIEW:REJECT]` |
|
|
420
520
|
instruction_template: |
|
|
421
521
|
## Workflow Context
|
|
422
|
-
- Iteration: {iteration}/{max_iterations}
|
|
522
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
523
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
423
524
|
- Step: ai_review (AI-Generated Code Review)
|
|
424
525
|
- Report Directory: .takt/reports/{report_dir}/
|
|
425
526
|
- Report File: .takt/reports/{report_dir}/05-ai-review.md
|
|
@@ -441,7 +542,7 @@ steps:
|
|
|
441
542
|
|
|
442
543
|
**Report output:** Output to the `Report File` specified above.
|
|
443
544
|
- If file does not exist: Create new file
|
|
444
|
-
- If file exists: Append with `## Iteration {
|
|
545
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
445
546
|
|
|
446
547
|
**Report format:**
|
|
447
548
|
```markdown
|
|
@@ -490,9 +591,22 @@ steps:
|
|
|
490
591
|
- Bash
|
|
491
592
|
- WebSearch
|
|
492
593
|
- WebFetch
|
|
594
|
+
status_rules_prompt: |
|
|
595
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
596
|
+
|
|
597
|
+
**Without this tag, the workflow will stop.**
|
|
598
|
+
Your final output MUST include a status tag following the rules below.
|
|
599
|
+
|
|
600
|
+
## Output Format
|
|
601
|
+
|
|
602
|
+
| Situation | Tag |
|
|
603
|
+
|-----------|-----|
|
|
604
|
+
| Fix complete | `[CODER:DONE]` |
|
|
605
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
493
606
|
instruction_template: |
|
|
494
607
|
## Workflow Context
|
|
495
|
-
- Iteration: {iteration}/{max_iterations}
|
|
608
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
609
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
496
610
|
- Step: ai_fix
|
|
497
611
|
|
|
498
612
|
## AI Review Feedback (This is the latest instruction - prioritize this)
|
|
@@ -532,9 +646,22 @@ steps:
|
|
|
532
646
|
- Grep
|
|
533
647
|
- WebSearch
|
|
534
648
|
- WebFetch
|
|
649
|
+
status_rules_prompt: |
|
|
650
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
651
|
+
|
|
652
|
+
**Without this tag, the workflow will stop.**
|
|
653
|
+
Your final output MUST include a status tag following the rules below.
|
|
654
|
+
|
|
655
|
+
## Output Format
|
|
656
|
+
|
|
657
|
+
| Situation | Tag |
|
|
658
|
+
|-----------|-----|
|
|
659
|
+
| No security issues | `[SECURITY:APPROVE]` |
|
|
660
|
+
| Vulnerabilities found | `[SECURITY:REJECT]` |
|
|
535
661
|
instruction_template: |
|
|
536
662
|
## Workflow Context
|
|
537
|
-
- Iteration: {iteration}/{max_iterations}
|
|
663
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
664
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
538
665
|
- Step: security_review (Security Expert Review)
|
|
539
666
|
- Report Directory: .takt/reports/{report_dir}/
|
|
540
667
|
- Report File: .takt/reports/{report_dir}/06-security-review.md
|
|
@@ -559,7 +686,7 @@ steps:
|
|
|
559
686
|
|
|
560
687
|
**Report output:** Output to the `Report File` specified above.
|
|
561
688
|
- If file does not exist: Create new file
|
|
562
|
-
- If file exists: Append with `## Iteration {
|
|
689
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
563
690
|
|
|
564
691
|
**Report format:**
|
|
565
692
|
```markdown
|
|
@@ -606,9 +733,23 @@ steps:
|
|
|
606
733
|
- Bash
|
|
607
734
|
- WebSearch
|
|
608
735
|
- WebFetch
|
|
736
|
+
status_rules_prompt: |
|
|
737
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
738
|
+
|
|
739
|
+
**Without this tag, the workflow will stop.**
|
|
740
|
+
Your final output MUST include a status tag following the rules below.
|
|
741
|
+
|
|
742
|
+
## Output Format
|
|
743
|
+
|
|
744
|
+
| Situation | Tag |
|
|
745
|
+
|-----------|-----|
|
|
746
|
+
| Minor fix complete | `[CODER:DONE]` |
|
|
747
|
+
| Major fix (restart from CQRS+ES) | `[CODER:REJECT]` |
|
|
748
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
609
749
|
instruction_template: |
|
|
610
750
|
## Workflow Context
|
|
611
|
-
- Iteration: {iteration}/{max_iterations}
|
|
751
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
752
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
612
753
|
- Step: fix_security
|
|
613
754
|
|
|
614
755
|
## Security Review Feedback (This is the latest instruction - prioritize this)
|
|
@@ -659,9 +800,22 @@ steps:
|
|
|
659
800
|
- Grep
|
|
660
801
|
- WebSearch
|
|
661
802
|
- WebFetch
|
|
803
|
+
status_rules_prompt: |
|
|
804
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
805
|
+
|
|
806
|
+
**Without this tag, the workflow will stop.**
|
|
807
|
+
Your final output MUST include a status tag following the rules below.
|
|
808
|
+
|
|
809
|
+
## Output Format
|
|
810
|
+
|
|
811
|
+
| Situation | Tag |
|
|
812
|
+
|-----------|-----|
|
|
813
|
+
| Quality standards met | `[QA:APPROVE]` |
|
|
814
|
+
| Quality issues found | `[QA:REJECT]` |
|
|
662
815
|
instruction_template: |
|
|
663
816
|
## Workflow Context
|
|
664
|
-
- Iteration: {iteration}/{max_iterations}
|
|
817
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
818
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
665
819
|
- Step: qa_review (QA Expert Review)
|
|
666
820
|
- Report Directory: .takt/reports/{report_dir}/
|
|
667
821
|
- Report File: .takt/reports/{report_dir}/07-qa-review.md
|
|
@@ -687,7 +841,7 @@ steps:
|
|
|
687
841
|
|
|
688
842
|
**Report output:** Output to the `Report File` specified above.
|
|
689
843
|
- If file does not exist: Create new file
|
|
690
|
-
- If file exists: Append with `## Iteration {
|
|
844
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
691
845
|
|
|
692
846
|
**Report format:**
|
|
693
847
|
```markdown
|
|
@@ -733,9 +887,24 @@ steps:
|
|
|
733
887
|
- Bash
|
|
734
888
|
- WebSearch
|
|
735
889
|
- WebFetch
|
|
890
|
+
status_rules_prompt: |
|
|
891
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
892
|
+
|
|
893
|
+
**Without this tag, the workflow will stop.**
|
|
894
|
+
Your final output MUST include a status tag following the rules below.
|
|
895
|
+
|
|
896
|
+
## Output Format
|
|
897
|
+
|
|
898
|
+
| Situation | Tag |
|
|
899
|
+
|-----------|-----|
|
|
900
|
+
| Minor fix complete | `[CODER:DONE]` |
|
|
901
|
+
| Security-impacting fix | `[CODER:IMPROVE]` |
|
|
902
|
+
| Major fix (restart from CQRS+ES) | `[CODER:REJECT]` |
|
|
903
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
736
904
|
instruction_template: |
|
|
737
905
|
## Workflow Context
|
|
738
|
-
- Iteration: {iteration}/{max_iterations}
|
|
906
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
907
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
739
908
|
- Step: fix_qa
|
|
740
909
|
|
|
741
910
|
## QA Review Feedback (This is the latest instruction - prioritize this)
|
|
@@ -790,9 +959,22 @@ steps:
|
|
|
790
959
|
- Grep
|
|
791
960
|
- WebSearch
|
|
792
961
|
- WebFetch
|
|
962
|
+
status_rules_prompt: |
|
|
963
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
964
|
+
|
|
965
|
+
**Without this tag, the workflow will stop.**
|
|
966
|
+
Your final output MUST include a status tag following the rules below.
|
|
967
|
+
|
|
968
|
+
## Output Format
|
|
969
|
+
|
|
970
|
+
| Situation | Tag |
|
|
971
|
+
|-----------|-----|
|
|
972
|
+
| Ready to merge | `[SUPERVISOR:APPROVE]` |
|
|
973
|
+
| Issues found | `[SUPERVISOR:REJECT]` |
|
|
793
974
|
instruction_template: |
|
|
794
975
|
## Workflow Context
|
|
795
|
-
- Iteration: {iteration}/{max_iterations}
|
|
976
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
977
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
796
978
|
- Step: supervise (Final Review)
|
|
797
979
|
- Report Directory: .takt/reports/{report_dir}/
|
|
798
980
|
- Report Files:
|
|
@@ -828,7 +1010,7 @@ steps:
|
|
|
828
1010
|
|
|
829
1011
|
**Report output:** Output to the `Report Files` specified above.
|
|
830
1012
|
- If file does not exist: Create new file
|
|
831
|
-
- If file exists: Append with `## Iteration {
|
|
1013
|
+
- If file exists: Append with `## Iteration {step_iteration}` section
|
|
832
1014
|
|
|
833
1015
|
**Validation report format:**
|
|
834
1016
|
```markdown
|
|
@@ -906,9 +1088,22 @@ steps:
|
|
|
906
1088
|
- Bash
|
|
907
1089
|
- WebSearch
|
|
908
1090
|
- WebFetch
|
|
1091
|
+
status_rules_prompt: |
|
|
1092
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
1093
|
+
|
|
1094
|
+
**Without this tag, the workflow will stop.**
|
|
1095
|
+
Your final output MUST include a status tag following the rules below.
|
|
1096
|
+
|
|
1097
|
+
## Output Format
|
|
1098
|
+
|
|
1099
|
+
| Situation | Tag |
|
|
1100
|
+
|-----------|-----|
|
|
1101
|
+
| Fix complete | `[CODER:DONE]` |
|
|
1102
|
+
| Cannot proceed | `[CODER:BLOCKED]` |
|
|
909
1103
|
instruction_template: |
|
|
910
1104
|
## Workflow Context
|
|
911
|
-
- Iteration: {iteration}/{max_iterations}
|
|
1105
|
+
- Iteration: {iteration}/{max_iterations} (workflow-wide)
|
|
1106
|
+
- Step Iteration: {step_iteration} (times this step has run)
|
|
912
1107
|
- Step: fix_supervisor
|
|
913
1108
|
|
|
914
1109
|
## Supervisor Feedback (This is the latest instruction - prioritize this)
|
|
@@ -1,6 +1,16 @@
|
|
|
1
1
|
# MAGI System Workflow
|
|
2
2
|
# A deliberation workflow modeled after Evangelion's MAGI system
|
|
3
3
|
# Three personas (scientist, nurturer, pragmatist) analyze from different perspectives and vote
|
|
4
|
+
#
|
|
5
|
+
# Template Variables:
|
|
6
|
+
# {iteration} - Workflow-wide turn count (total steps executed across all agents)
|
|
7
|
+
# {max_iterations} - Maximum iterations allowed for the workflow
|
|
8
|
+
# {step_iteration} - Per-step iteration count (how many times THIS step has been executed)
|
|
9
|
+
# {task} - Original user request
|
|
10
|
+
# {previous_response} - Output from the previous step
|
|
11
|
+
# {git_diff} - Current uncommitted changes (git diff)
|
|
12
|
+
# {user_inputs} - Accumulated user inputs during workflow
|
|
13
|
+
# {report_dir} - Report directory name (e.g., "20250126-143052-task-summary")
|
|
4
14
|
|
|
5
15
|
name: magi
|
|
6
16
|
description: MAGI Deliberation System - Analyze from 3 perspectives and decide by majority
|
|
@@ -16,6 +26,27 @@ steps:
|
|
|
16
26
|
- Grep
|
|
17
27
|
- WebSearch
|
|
18
28
|
- WebFetch
|
|
29
|
+
status_rules_prompt: |
|
|
30
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
31
|
+
|
|
32
|
+
**Without this tag, the workflow will stop.**
|
|
33
|
+
Your final output MUST include a status tag following the rules below.
|
|
34
|
+
|
|
35
|
+
## Output Format
|
|
36
|
+
|
|
37
|
+
| Judgment | Tag |
|
|
38
|
+
|----------|-----|
|
|
39
|
+
| In favor | `[MELCHIOR:APPROVE]` |
|
|
40
|
+
| Against | `[MELCHIOR:REJECT]` |
|
|
41
|
+
| Conditional approval | `[MELCHIOR:CONDITIONAL]` |
|
|
42
|
+
|
|
43
|
+
### Output Example
|
|
44
|
+
|
|
45
|
+
```
|
|
46
|
+
[MELCHIOR:APPROVE]
|
|
47
|
+
|
|
48
|
+
Reason: {Reason for approval}
|
|
49
|
+
```
|
|
19
50
|
instruction_template: |
|
|
20
51
|
# MAGI System Initiated
|
|
21
52
|
|
|
@@ -25,11 +56,6 @@ steps:
|
|
|
25
56
|
## Instructions
|
|
26
57
|
You are MELCHIOR-1 of the MAGI System.
|
|
27
58
|
Analyze the above from the perspective of a scientist/engineer and render your judgment.
|
|
28
|
-
|
|
29
|
-
Your judgment must be one of:
|
|
30
|
-
- [MELCHIOR:APPROVE] - In favor
|
|
31
|
-
- [MELCHIOR:REJECT] - Against
|
|
32
|
-
- [MELCHIOR:CONDITIONAL] - Conditional approval
|
|
33
59
|
transitions:
|
|
34
60
|
- condition: always
|
|
35
61
|
next_step: balthasar
|
|
@@ -42,6 +68,27 @@ steps:
|
|
|
42
68
|
- Grep
|
|
43
69
|
- WebSearch
|
|
44
70
|
- WebFetch
|
|
71
|
+
status_rules_prompt: |
|
|
72
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
73
|
+
|
|
74
|
+
**Without this tag, the workflow will stop.**
|
|
75
|
+
Your final output MUST include a status tag following the rules below.
|
|
76
|
+
|
|
77
|
+
## Output Format
|
|
78
|
+
|
|
79
|
+
| Judgment | Tag |
|
|
80
|
+
|----------|-----|
|
|
81
|
+
| In favor | `[BALTHASAR:APPROVE]` |
|
|
82
|
+
| Against | `[BALTHASAR:REJECT]` |
|
|
83
|
+
| Conditional approval | `[BALTHASAR:CONDITIONAL]` |
|
|
84
|
+
|
|
85
|
+
### Output Example
|
|
86
|
+
|
|
87
|
+
```
|
|
88
|
+
[BALTHASAR:APPROVE]
|
|
89
|
+
|
|
90
|
+
Reason: {Reason for approval}
|
|
91
|
+
```
|
|
45
92
|
instruction_template: |
|
|
46
93
|
# MAGI System Continuing
|
|
47
94
|
|
|
@@ -55,11 +102,6 @@ steps:
|
|
|
55
102
|
You are BALTHASAR-2 of the MAGI System.
|
|
56
103
|
Analyze the above from the perspective of a nurturer and render your judgment.
|
|
57
104
|
Consider MELCHIOR's judgment as reference, but make your own independent assessment.
|
|
58
|
-
|
|
59
|
-
Your judgment must be one of:
|
|
60
|
-
- [BALTHASAR:APPROVE] - In favor
|
|
61
|
-
- [BALTHASAR:REJECT] - Against
|
|
62
|
-
- [BALTHASAR:CONDITIONAL] - Conditional approval
|
|
63
105
|
pass_previous_response: true
|
|
64
106
|
transitions:
|
|
65
107
|
- condition: always
|
|
@@ -73,28 +115,24 @@ steps:
|
|
|
73
115
|
- Grep
|
|
74
116
|
- WebSearch
|
|
75
117
|
- WebFetch
|
|
76
|
-
|
|
77
|
-
#
|
|
118
|
+
status_rules_prompt: |
|
|
119
|
+
# ⚠️ REQUIRED: Status Output Rules ⚠️
|
|
78
120
|
|
|
79
|
-
|
|
80
|
-
|
|
121
|
+
**Without this tag, the workflow will stop.**
|
|
122
|
+
Your final output MUST include a status tag following the rules below.
|
|
81
123
|
|
|
82
|
-
##
|
|
83
|
-
{previous_response}
|
|
124
|
+
## Output Format
|
|
84
125
|
|
|
85
|
-
|
|
86
|
-
You are CASPER-3 of the MAGI System.
|
|
87
|
-
Analyze the above from a practical/realistic perspective and render your judgment.
|
|
126
|
+
Final conclusion by majority vote:
|
|
88
127
|
|
|
89
|
-
|
|
128
|
+
| Judgment | Tag |
|
|
129
|
+
|----------|-----|
|
|
130
|
+
| Approved (2+ in favor) | `[MAGI:APPROVE]` |
|
|
131
|
+
| Rejected (2+ against) | `[MAGI:REJECT]` |
|
|
132
|
+
| Conditional approval | `[MAGI:CONDITIONAL]` |
|
|
90
133
|
|
|
91
|
-
###
|
|
92
|
-
Determine the final judgment by majority vote:
|
|
93
|
-
- [MAGI:APPROVE] - Approved (2 or more in favor)
|
|
94
|
-
- [MAGI:REJECT] - Rejected (2 or more against)
|
|
95
|
-
- [MAGI:CONDITIONAL] - Conditional approval (majority conditional or split opinions)
|
|
134
|
+
### Output Example
|
|
96
135
|
|
|
97
|
-
**Final Conclusion Format Example:**
|
|
98
136
|
```
|
|
99
137
|
## MAGI System Final Judgment
|
|
100
138
|
|
|
@@ -108,6 +146,20 @@ steps:
|
|
|
108
146
|
|
|
109
147
|
[Reasoning/Summary]
|
|
110
148
|
```
|
|
149
|
+
instruction_template: |
|
|
150
|
+
# MAGI System Final Deliberation
|
|
151
|
+
|
|
152
|
+
## Matter for Deliberation
|
|
153
|
+
{task}
|
|
154
|
+
|
|
155
|
+
## Previous Judgments
|
|
156
|
+
{previous_response}
|
|
157
|
+
|
|
158
|
+
## Instructions
|
|
159
|
+
You are CASPER-3 of the MAGI System.
|
|
160
|
+
Analyze the above from a practical/realistic perspective and render your judgment.
|
|
161
|
+
|
|
162
|
+
**Finally, tally the judgments from all three and provide the final conclusion.**
|
|
111
163
|
pass_previous_response: true
|
|
112
164
|
transitions:
|
|
113
165
|
- condition: always
|