siesa-agents 2.1.17-dev.5 → 2.1.17-dev.6

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.
Files changed (3) hide show
  1. package/bmad-core/tasks/create-doc.md +54 -10
  2. package/claude/commands/BMad/tasks/create-doc.md +50 -10
  3. package/package.json +1 -1
package/bmad-core/tasks/create-doc.md CHANGED
@@ -1,3 +1,7 @@
1
+ # /create-doc Task
2
+
3
+ When this command is used, execute the following task:
4
+
1
5
  <!-- Powered by BMAD™ Core -->
2
6
 
3
7
  # Create Document from Template (YAML Driven)
@@ -40,15 +44,25 @@ If a YAML Template has not been provided, list all templates from .bmad-core/tem
40
44
  ## Processing Flow
41
45
 
42
46
  1. **Parse YAML template** - Load template metadata and sections
43
- 2. **Set preferences** - Show current mode (Interactive), confirm output file
44
- 3. **Process each section:**
45
- - Skip if condition unmet
46
- - Check agent permissions (owner/editors) - note if section is restricted to specific agents
47
- - Draft content using section instruction
48
- - Present content + detailed rationale
49
- - **IF elicit: true** → MANDATORY 1-9 options format
50
- - Save to file if possible
51
- 4. **Continue until complete**
47
+ 2. **Set preferences** - Show current mode (Interactive or YOLO), confirm output file
48
+ 3. **If YOLO mode:**
49
+ - Count total sections
50
+ - Estimate optimal batch count based on section complexity
51
+ - Split document into 3-4 batches (~12-17 sections each)
52
+ - Process each batch → Save → STOP → Wait for "continue"
53
+ - Repeat until all batches complete
54
+
55
+ 4. **If Interactive mode:**
56
+ - Process each section one at a time:
57
+ - Skip if condition unmet
58
+ - Check agent permissions (owner/editors)
59
+ - Draft content using section instruction
60
+ - Present content + detailed rationale
61
+ - **IF elicit: true** → MANDATORY 1-9 options format
62
+ - Save to file after section
63
+ - Wait for user response before next section
64
+
65
+ 5. **Continue until all sections complete**
52
66
 
53
67
  ## Detailed Rationale Requirements
54
68
 
@@ -85,7 +99,37 @@ When processing sections with agent permission fields:
85
99
 
86
100
  ## YOLO Mode
87
101
 
88
- User can type `#yolo` to toggle to YOLO mode (process all sections at once).
102
+ User can type `#yolo` to toggle to YOLO mode (process sections in batches without elicitation).
103
+
104
+ **MANDATORY BATCH WORKFLOW:**
105
+
106
+ 1. **Automatic batch calculation:**
107
+ - Estimate: `sections × 600 tokens`
108
+ - Max per batch: **22,000 tokens** (safe 10K buffer - Claude can't measure in real-time)
109
+ - Batches needed: `total_estimate ÷ 22,000` (rounded up)
110
+ - If complex sections detected: multiply estimate × 1.5
111
+
112
+ 2. **CRITICAL - Repeatable sections (epics, stories, etc.):**
113
+ - Count total repeatable items (e.g., 10 epics × 10 stories each)
114
+ - Estimate tokens per item: `epic ~7,000 tokens`, `story ~800 tokens`
115
+ - Calculate: `items × tokens_per_item = total_tokens`
116
+ - If `total_tokens > 22,000`: Split into sub-batches
117
+ - Example: 10 epics × 7K = 70K tokens → 4 sub-batches (3-3-2-2 epics)
118
+ - Process each sub-batch separately with pauses
119
+
120
+ 3. **Process EACH batch:**
121
+ - Generate content for batch sections
122
+ - Save to file
123
+ - STOP and tell user: "✅ Batch X/Y complete: Sections A-B saved. Type 'continue'."
124
+ - WAIT for user to type "continue"
125
+ - Repeat until all batches complete
126
+
127
+ **Examples:**
128
+ - 30 sections (~18K tokens) → 1 batch (0 pauses)
129
+ - 50 sections (~30K tokens) → 2 batches (1 pause)
130
+ - 50 sections + 10 epics (~80K tokens) → 4 batches (3 pauses)
131
+
132
+ **THIS IS NON-NEGOTIABLE FOR YOLO MODE**
89
133
 
90
134
  ## CRITICAL REMINDERS
91
135
 
package/claude/commands/BMad/tasks/create-doc.md CHANGED
@@ -44,15 +44,25 @@ If a YAML Template has not been provided, list all templates from .bmad-core/tem
44
44
  ## Processing Flow
45
45
 
46
46
  1. **Parse YAML template** - Load template metadata and sections
47
- 2. **Set preferences** - Show current mode (Interactive), confirm output file
48
- 3. **Process each section:**
49
- - Skip if condition unmet
50
- - Check agent permissions (owner/editors) - note if section is restricted to specific agents
51
- - Draft content using section instruction
52
- - Present content + detailed rationale
53
- - **IF elicit: true** → MANDATORY 1-9 options format
54
- - Save to file if possible
55
- 4. **Continue until complete**
47
+ 2. **Set preferences** - Show current mode (Interactive or YOLO), confirm output file
48
+ 3. **If YOLO mode:**
49
+ - Count total sections
50
+ - Estimate optimal batch count based on section complexity
51
+ - Split document into 3-4 batches (~12-17 sections each)
52
+ - Process each batch → Save → STOP → Wait for "continue"
53
+ - Repeat until all batches complete
54
+
55
+ 4. **If Interactive mode:**
56
+ - Process each section one at a time:
57
+ - Skip if condition unmet
58
+ - Check agent permissions (owner/editors)
59
+ - Draft content using section instruction
60
+ - Present content + detailed rationale
61
+ - **IF elicit: true** → MANDATORY 1-9 options format
62
+ - Save to file after section
63
+ - Wait for user response before next section
64
+
65
+ 5. **Continue until all sections complete**
56
66
 
57
67
  ## Detailed Rationale Requirements
58
68
 
@@ -89,7 +99,37 @@ When processing sections with agent permission fields:
89
99
 
90
100
  ## YOLO Mode
91
101
 
92
- User can type `#yolo` to toggle to YOLO mode (process all sections at once).
102
+ User can type `#yolo` to toggle to YOLO mode (process sections in batches without elicitation).
103
+
104
+ **MANDATORY BATCH WORKFLOW:**
105
+
106
+ 1. **Automatic batch calculation:**
107
+ - Estimate: `sections × 600 tokens`
108
+ - Max per batch: **22,000 tokens** (safe 10K buffer - Claude can't measure in real-time)
109
+ - Batches needed: `total_estimate ÷ 22,000` (rounded up)
110
+ - If complex sections detected: multiply estimate × 1.5
111
+
112
+ 2. **CRITICAL - Repeatable sections (epics, stories, etc.):**
113
+ - Count total repeatable items (e.g., 10 epics × 10 stories each)
114
+ - Estimate tokens per item: `epic ~7,000 tokens`, `story ~800 tokens`
115
+ - Calculate: `items × tokens_per_item = total_tokens`
116
+ - If `total_tokens > 22,000`: Split into sub-batches
117
+ - Example: 10 epics × 7K = 70K tokens → 4 sub-batches (3-3-2-2 epics)
118
+ - Process each sub-batch separately with pauses
119
+
120
+ 3. **Process EACH batch:**
121
+ - Generate content for batch sections
122
+ - Save to file
123
+ - STOP and tell user: "✅ Batch X/Y complete: Sections A-B saved. Type 'continue'."
124
+ - WAIT for user to type "continue"
125
+ - Repeat until all batches complete
126
+
127
+ **Examples:**
128
+ - 30 sections (~18K tokens) → 1 batch (0 pauses)
129
+ - 50 sections (~30K tokens) → 2 batches (1 pause)
130
+ - 50 sections + 10 epics (~80K tokens) → 4 batches (3 pauses)
131
+
132
+ **THIS IS NON-NEGOTIABLE FOR YOLO MODE**
93
133
 
94
134
  ## CRITICAL REMINDERS
95
135
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "siesa-agents",
3
- "version": "2.1.17-dev.5",
3
+ "version": "2.1.17-dev.6",
4
4
  "description": "Paquete para instalar y configurar agentes SIESA en tu proyecto",
5
5
  "main": "index.js",
6
6
  "bin": {