@e0ipso/ai-task-manager 1.26.4 → 1.26.5
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/package.json +1 -1
- package/templates/assistant/commands/tasks/create-plan.md +46 -27
- package/templates/assistant/commands/tasks/execute-blueprint.md +46 -27
- package/templates/assistant/commands/tasks/execute-task.md +43 -24
- package/templates/assistant/commands/tasks/fix-broken-tests.md +1 -13
- package/templates/assistant/commands/tasks/full-workflow.md +57 -38
- package/templates/assistant/commands/tasks/generate-tasks.md +44 -27
- package/templates/assistant/commands/tasks/refine-plan.md +46 -27
package/package.json
CHANGED
|
@@ -8,25 +8,55 @@ You are a strategic planning specialist who creates actionable plan documents th
|
|
|
8
8
|
disciplined scope control. Your role is to think hard to create detailed, actionable plans based on user input while
|
|
9
9
|
ensuring you have all necessary context before proceeding. Use the plan-creator sub-agent for this if it is available.
|
|
10
10
|
|
|
11
|
-
|
|
12
|
-
|
|
13
|
-
Before proceeding with this command, you MUST load and respect the assistant's configuration:
|
|
11
|
+
---
|
|
14
12
|
|
|
15
|
-
|
|
16
|
-
1. `/workspace/AGENTS.md` - Project-level task management guidance
|
|
17
|
-
2. `/workspace/CLAUDE.md` - Claude-specific assistant configuration (if it exists)
|
|
18
|
-
3. `/home/node/.claude/CLAUDE.md` - Global Claude configuration from your home directory (if it exists)
|
|
13
|
+
Ultrathink, think harder, and use tools.
|
|
19
14
|
|
|
20
|
-
|
|
15
|
+
## Find the AI Task Manager root
|
|
21
16
|
|
|
22
|
-
|
|
17
|
+
```bash
|
|
18
|
+
if [ ! -f /tmp/find-ai-task-manager-root.js ]; then
|
|
19
|
+
cat << 'EOF' > /tmp/find-ai-task-manager-root.js
|
|
20
|
+
const fs = require('fs');
|
|
21
|
+
const path = require('path');
|
|
22
|
+
|
|
23
|
+
const findRoot = (currentDir) => {
|
|
24
|
+
const taskManagerPath = path.join(currentDir, '.ai/task-manager');
|
|
25
|
+
const metadataPath = path.join(taskManagerPath, '.init-metadata.json');
|
|
26
|
+
|
|
27
|
+
try {
|
|
28
|
+
if (fs.existsSync(metadataPath) && JSON.parse(fs.readFileSync(metadataPath, 'utf8')).version) {
|
|
29
|
+
console.log(path.resolve(taskManagerPath));
|
|
30
|
+
process.exit(0);
|
|
31
|
+
}
|
|
32
|
+
} catch (e) {
|
|
33
|
+
// Continue searching
|
|
34
|
+
}
|
|
35
|
+
|
|
36
|
+
const parentDir = path.dirname(currentDir);
|
|
37
|
+
if (parentDir !== currentDir) {
|
|
38
|
+
findRoot(parentDir);
|
|
39
|
+
} else {
|
|
40
|
+
process.exit(1);
|
|
41
|
+
}
|
|
42
|
+
};
|
|
43
|
+
|
|
44
|
+
findRoot(process.cwd());
|
|
45
|
+
EOF
|
|
46
|
+
fi
|
|
23
47
|
|
|
24
|
-
|
|
48
|
+
root=$(node /tmp/find-ai-task-manager-root.js)
|
|
25
49
|
|
|
26
|
-
|
|
50
|
+
if [ -z "$root" ]; then
|
|
51
|
+
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
52
|
+
exit 1
|
|
53
|
+
fi
|
|
54
|
+
```
|
|
27
55
|
|
|
28
56
|
## Instructions
|
|
29
57
|
|
|
58
|
+
Include $root/ $root/.ai/task-manager/config/TASK_MANAGER.md for the directory structure of tasks.
|
|
59
|
+
|
|
30
60
|
The user input is:
|
|
31
61
|
|
|
32
62
|
<user-input>
|
|
@@ -39,11 +69,11 @@ If no user input is provided stop immediately and show an error message to the u
|
|
|
39
69
|
|
|
40
70
|
Use your internal Todo task tool to track the following plan generation:
|
|
41
71
|
|
|
42
|
-
- [ ] Read and execute
|
|
72
|
+
- [ ] Read and execute $root/ $root/.ai/task-manager/config/hooks/PRE_PLAN.md
|
|
43
73
|
- [ ] User input and context analysis
|
|
44
74
|
- [ ] Clarification questions
|
|
45
75
|
- [ ] Plan generation
|
|
46
|
-
- [ ] Read and execute
|
|
76
|
+
- [ ] Read and execute $root/ $root/.ai/task-manager/config/hooks/POST_PLAN.md
|
|
47
77
|
|
|
48
78
|
#### Step 1: Context Analysis
|
|
49
79
|
Before creating any plan, analyze the user's request for:
|
|
@@ -66,7 +96,7 @@ Try to answer your own questions first by inspecting the codebase, docs, and ass
|
|
|
66
96
|
IMPORTANT: Once you have the user's answers go back to Step 2. Do this in a loop until you have no more questions. Ask as many rounds of questions as necessary, it is very important you have all the information you need to achieve your task.
|
|
67
97
|
|
|
68
98
|
#### Step 3: Plan Generation
|
|
69
|
-
Only after confirming sufficient context, create a plan according the the
|
|
99
|
+
Only after confirming sufficient context, create a plan according the the $root/ $root/.ai/task-manager/config/templates/PLAN_TEMPLATE.md
|
|
70
100
|
|
|
71
101
|
##### CRITICAL: Output Format
|
|
72
102
|
|
|
@@ -88,7 +118,7 @@ This structured output enables automated workflow coordination and must be inclu
|
|
|
88
118
|
|
|
89
119
|
###### Plan Template
|
|
90
120
|
|
|
91
|
-
Use the template in
|
|
121
|
+
Use the template in $root/ $root/.ai/task-manager/config/templates/PLAN_TEMPLATE.md
|
|
92
122
|
|
|
93
123
|
###### Patterns to Avoid
|
|
94
124
|
Do not include the following in your plan output.
|
|
@@ -133,18 +163,7 @@ The schema for this frontmatter is:
|
|
|
133
163
|
|
|
134
164
|
### Plan ID Generation
|
|
135
165
|
|
|
136
|
-
|
|
137
|
-
|
|
138
|
-
```bash
|
|
139
|
-
root=$(node -e 'const fs=require("fs"),path=require("path");const f=p=>{const t=path.join(p,".ai/task-manager");const m=path.join(t,".init-metadata.json");try{if(JSON.parse(fs.readFileSync(m)).version){console.log(path.resolve(t));process.exit(0)}}catch(e){};const d=path.dirname(p);if(d!==p)f(d)};f(process.cwd());process.exit(1)')
|
|
140
|
-
|
|
141
|
-
if [ -z "$root" ]; then
|
|
142
|
-
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
143
|
-
exit 1
|
|
144
|
-
fi
|
|
145
|
-
```
|
|
146
|
-
|
|
147
|
-
Then execute this script to determine the plan ID:
|
|
166
|
+
Execute this script to determine the plan ID:
|
|
148
167
|
|
|
149
168
|
```bash
|
|
150
169
|
next_id=$(node $root/config/scripts/get-next-plan-id.cjs)
|
|
@@ -4,17 +4,6 @@ description: Execute the task in the plan.
|
|
|
4
4
|
---
|
|
5
5
|
# Task Execution
|
|
6
6
|
|
|
7
|
-
## Assistant Configuration
|
|
8
|
-
|
|
9
|
-
Before proceeding with this command, you MUST load and respect the assistant's configuration:
|
|
10
|
-
|
|
11
|
-
Load the following configuration files in order of precedence (later files override earlier ones):
|
|
12
|
-
1. `/workspace/AGENTS.md` - Project-level task management guidance
|
|
13
|
-
2. `/workspace/CLAUDE.md` - Claude-specific assistant configuration (if it exists)
|
|
14
|
-
3. `/home/node/.claude/CLAUDE.md` - Global Claude configuration from your home directory (if it exists)
|
|
15
|
-
|
|
16
|
-
These files contain your global and project-level configuration rules. You MUST keep these rules and guidelines in mind during all subsequent operations in this command.
|
|
17
|
-
|
|
18
7
|
---
|
|
19
8
|
|
|
20
9
|
You are the coordinator responsible for executing all tasks defined in the execution blueprint of a plan document, so choose an appropriate sub-agent for this role. Your role is to coordinate phase-by-phase execution, manage parallel task processing, and ensure validation gates pass before phase transitions.
|
|
@@ -47,7 +36,37 @@ Before proceeding with execution, validate that tasks exist and the execution bl
|
|
|
47
36
|
First, discover the task manager root directory:
|
|
48
37
|
|
|
49
38
|
```bash
|
|
50
|
-
|
|
39
|
+
if [ ! -f /tmp/find-ai-task-manager-root.cjs ]; then
|
|
40
|
+
cat << 'EOF' > /tmp/find-ai-task-manager-root.cjs
|
|
41
|
+
const fs = require('fs');
|
|
42
|
+
const path = require('path');
|
|
43
|
+
|
|
44
|
+
const findRoot = (currentDir) => {
|
|
45
|
+
const taskManagerPath = path.join(currentDir, '.ai/task-manager');
|
|
46
|
+
const metadataPath = path.join(taskManagerPath, '.init-metadata.json');
|
|
47
|
+
|
|
48
|
+
try {
|
|
49
|
+
if (fs.existsSync(metadataPath) && JSON.parse(fs.readFileSync(metadataPath, 'utf8')).version) {
|
|
50
|
+
console.log(path.resolve(taskManagerPath));
|
|
51
|
+
process.exit(0);
|
|
52
|
+
}
|
|
53
|
+
} catch (e) {
|
|
54
|
+
// Continue searching
|
|
55
|
+
}
|
|
56
|
+
|
|
57
|
+
const parentDir = path.dirname(currentDir);
|
|
58
|
+
if (parentDir !== currentDir) {
|
|
59
|
+
findRoot(parentDir);
|
|
60
|
+
} else {
|
|
61
|
+
process.exit(1);
|
|
62
|
+
}
|
|
63
|
+
};
|
|
64
|
+
|
|
65
|
+
findRoot(process.cwd());
|
|
66
|
+
EOF
|
|
67
|
+
fi
|
|
68
|
+
|
|
69
|
+
root=$(node /tmp/find-ai-task-manager-root.js)
|
|
51
70
|
|
|
52
71
|
if [ -z "$root" ]; then
|
|
53
72
|
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
@@ -95,21 +114,21 @@ Use your internal Todo task tool to track the execution of all phases, and the f
|
|
|
95
114
|
|
|
96
115
|
- [ ] Create feature branch from the main branch.
|
|
97
116
|
- [ ] Validate or auto-generate tasks and execution blueprint if missing.
|
|
98
|
-
- [ ] Execute
|
|
117
|
+
- [ ] Execute $root/.ai/task-manager/config/hooks/PRE_PHASE.md hook before Phase 1.
|
|
99
118
|
- [ ] Phase 1: Execute 1 task(s) in parallel.
|
|
100
|
-
- [ ] Execute
|
|
101
|
-
- [ ] Execute
|
|
119
|
+
- [ ] Execute $root/.ai/task-manager/config/hooks/POST_PHASE.md hook after Phase 1.
|
|
120
|
+
- [ ] Execute $root/.ai/task-manager/config/hooks/PRE_PHASE.md hook before Phase 2.
|
|
102
121
|
- [ ] Phase 2: Execute 3 task(s) in parallel.
|
|
103
|
-
- [ ] Execute
|
|
104
|
-
- [ ] Execute
|
|
122
|
+
- [ ] Execute $root/.ai/task-manager/config/hooks/POST_PHASE.md hook after Phase 2.
|
|
123
|
+
- [ ] Execute $root/.ai/task-manager/config/hooks/PRE_PHASE.md hook before Phase 3.
|
|
105
124
|
- [ ] Phase 3: Execute 1 task(s) in parallel.
|
|
106
|
-
- [ ] Execute
|
|
107
|
-
- [ ] Update the Plan 7 with execution summary using
|
|
125
|
+
- [ ] Execute $root/.ai/task-manager/config/hooks/POST_PHASE.md hook after Phase 3.
|
|
126
|
+
- [ ] Update the Plan 7 with execution summary using $root/.ai/task-manager/config/hooks/EXECUTION_SUMMARY_TEMPLATE.md.
|
|
108
127
|
- [ ] Archive Plan 7.
|
|
109
128
|
|
|
110
129
|
### Phase Pre-Execution
|
|
111
130
|
|
|
112
|
-
Read and execute
|
|
131
|
+
Read and execute $root/.ai/task-manager/config/hooks/PRE_PHASE.md
|
|
113
132
|
|
|
114
133
|
### Phase Execution Workflow
|
|
115
134
|
|
|
@@ -118,7 +137,7 @@ Read and execute .ai/task-manager/config/hooks/PRE_PHASE.md
|
|
|
118
137
|
- List all tasks scheduled for parallel execution in this phase
|
|
119
138
|
|
|
120
139
|
2. **Agent Selection and Task Assignment**
|
|
121
|
-
Read and execute
|
|
140
|
+
Read and execute $root/.ai/task-manager/config/hooks/PRE_TASK_ASSIGNMENT.md
|
|
122
141
|
|
|
123
142
|
3. **Parallel Execution**
|
|
124
143
|
- Deploy all selected agents simultaneously using your internal Task tool
|
|
@@ -133,7 +152,7 @@ Read and execute .ai/task-manager/config/hooks/PRE_TASK_ASSIGNMENT.md
|
|
|
133
152
|
|
|
134
153
|
### Phase Post-Execution
|
|
135
154
|
|
|
136
|
-
Read and execute
|
|
155
|
+
Read and execute $root/.ai/task-manager/config/hooks/POST_PHASE.md
|
|
137
156
|
|
|
138
157
|
|
|
139
158
|
### Phase Transition
|
|
@@ -145,14 +164,14 @@ Read and execute .ai/task-manager/config/hooks/POST_PHASE.md
|
|
|
145
164
|
### Error Handling
|
|
146
165
|
|
|
147
166
|
#### Validation Gate Failures
|
|
148
|
-
Read and execute
|
|
167
|
+
Read and execute $root/.ai/task-manager/config/hooks/POST_ERROR_DETECTION.md
|
|
149
168
|
|
|
150
169
|
### Output Requirements
|
|
151
170
|
|
|
152
171
|
**Output Behavior:**
|
|
153
172
|
|
|
154
173
|
Provide a concise execution summary:
|
|
155
|
-
- Example: "Execution completed. Review summary:
|
|
174
|
+
- Example: "Execution completed. Review summary: `$root/.ai/task-manager/archive/[plan]/plan-[id].md`"
|
|
156
175
|
|
|
157
176
|
**CRITICAL - Structured Output for Command Coordination:**
|
|
158
177
|
|
|
@@ -163,7 +182,7 @@ Always end your output with a standardized summary in this exact format:
|
|
|
163
182
|
Execution Summary:
|
|
164
183
|
- Plan ID: [numeric-id]
|
|
165
184
|
- Status: Archived
|
|
166
|
-
- Location:
|
|
185
|
+
- Location: $root/.ai/task-manager/archive/[plan-id]--[plan-name]/
|
|
167
186
|
```
|
|
168
187
|
|
|
169
188
|
This structured output enables automated workflow coordination and must be included even when running standalone.
|
|
@@ -184,7 +203,7 @@ Upon successful completion of all phases and validation gates, perform the follo
|
|
|
184
203
|
|
|
185
204
|
### 1. Execution Summary Generation
|
|
186
205
|
|
|
187
|
-
Append an execution summary section to the plan document with the format described in
|
|
206
|
+
Append an execution summary section to the plan document with the format described in $root/.ai/task-manager/config/templates/[EXECUTION_SUMMARY_TEMPLATE.md
|
|
188
207
|
|
|
189
208
|
### 2. Plan Archival
|
|
190
209
|
|
|
@@ -192,7 +211,7 @@ After successfully appending the execution summary:
|
|
|
192
211
|
|
|
193
212
|
**Move completed plan to archive**:
|
|
194
213
|
```bash
|
|
195
|
-
mv
|
|
214
|
+
mv $root/.ai/task-manager/plans/[plan-folder] $root/.ai/task-manager/archive/
|
|
196
215
|
```
|
|
197
216
|
|
|
198
217
|
### Important Notes
|
|
@@ -6,18 +6,48 @@ description: Execute a single task with dependency validation and status managem
|
|
|
6
6
|
|
|
7
7
|
You are responsible for executing a single task within a plan while maintaining strict dependency validation and proper status management. Your role is to ensure the task is ready for execution, deploy the appropriate agent, and track execution progress.
|
|
8
8
|
|
|
9
|
-
|
|
9
|
+
---
|
|
10
10
|
|
|
11
|
-
|
|
11
|
+
## Find the AI Task Manager root
|
|
12
12
|
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
13
|
+
```bash
|
|
14
|
+
if [ ! -f /tmp/find-ai-task-manager-root.js ]; then
|
|
15
|
+
cat << 'EOF' > /tmp/find-ai-task-manager-root.js
|
|
16
|
+
const fs = require('fs');
|
|
17
|
+
const path = require('path');
|
|
18
|
+
|
|
19
|
+
const findRoot = (currentDir) => {
|
|
20
|
+
const taskManagerPath = path.join(currentDir, '.ai/task-manager');
|
|
21
|
+
const metadataPath = path.join(taskManagerPath, '.init-metadata.json');
|
|
22
|
+
|
|
23
|
+
try {
|
|
24
|
+
if (fs.existsSync(metadataPath) && JSON.parse(fs.readFileSync(metadataPath, 'utf8')).version) {
|
|
25
|
+
console.log(path.resolve(taskManagerPath));
|
|
26
|
+
process.exit(0);
|
|
27
|
+
}
|
|
28
|
+
} catch (e) {
|
|
29
|
+
// Continue searching
|
|
30
|
+
}
|
|
31
|
+
|
|
32
|
+
const parentDir = path.dirname(currentDir);
|
|
33
|
+
if (parentDir !== currentDir) {
|
|
34
|
+
findRoot(parentDir);
|
|
35
|
+
} else {
|
|
36
|
+
process.exit(1);
|
|
37
|
+
}
|
|
38
|
+
};
|
|
39
|
+
|
|
40
|
+
findRoot(process.cwd());
|
|
41
|
+
EOF
|
|
42
|
+
fi
|
|
17
43
|
|
|
18
|
-
|
|
44
|
+
root=$(node /tmp/find-ai-task-manager-root.js)
|
|
19
45
|
|
|
20
|
-
|
|
46
|
+
if [ -z "$root" ]; then
|
|
47
|
+
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
48
|
+
exit 1
|
|
49
|
+
fi
|
|
50
|
+
```
|
|
21
51
|
|
|
22
52
|
Use your internal Todo task tool to track the execution of all parts of the task, and the final update of noteworthy items during execution. Example:
|
|
23
53
|
|
|
@@ -40,7 +70,7 @@ Use your internal Todo task tool to track the execution of all parts of the task
|
|
|
40
70
|
- Plan ID: $1 (required)
|
|
41
71
|
- Task ID: $2 (required)
|
|
42
72
|
- Task management directory structure: `/`
|
|
43
|
-
- Dependency checking script:
|
|
73
|
+
- Dependency checking script: `$root/.ai/task-manager/config/scripts/check-task-dependencies.cjs`
|
|
44
74
|
|
|
45
75
|
### Input Validation
|
|
46
76
|
|
|
@@ -57,20 +87,9 @@ fi
|
|
|
57
87
|
|
|
58
88
|
## Execution Process
|
|
59
89
|
|
|
60
|
-
### 1.
|
|
61
|
-
|
|
62
|
-
First, discover the task manager root directory:
|
|
63
|
-
|
|
64
|
-
```bash
|
|
65
|
-
root=$(node -e 'const fs=require("fs"),path=require("path");const f=p=>{const t=path.join(p,".ai/task-manager");const m=path.join(t,".init-metadata.json");try{if(JSON.parse(fs.readFileSync(m)).version){console.log(path.resolve(t));process.exit(0)}}catch(e){};const d=path.dirname(p);if(d!==p)f(d)};f(process.cwd());process.exit(1)')
|
|
66
|
-
|
|
67
|
-
if [ -z "$root" ]; then
|
|
68
|
-
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
69
|
-
exit 1
|
|
70
|
-
fi
|
|
71
|
-
```
|
|
90
|
+
### 1. Plan Location
|
|
72
91
|
|
|
73
|
-
|
|
92
|
+
Locate the plan directory using the discovered root:
|
|
74
93
|
|
|
75
94
|
```bash
|
|
76
95
|
plan_id="$1"
|
|
@@ -173,7 +192,7 @@ echo "✓ All dependencies resolved - proceeding with execution"
|
|
|
173
192
|
|
|
174
193
|
Read task skills and select appropriate task-specific agent:
|
|
175
194
|
|
|
176
|
-
Read and execute
|
|
195
|
+
Read and execute $root/.ai/task-manager/config/hooks/PRE_TASK_ASSIGNMENT.md
|
|
177
196
|
|
|
178
197
|
### 6. Status Update to In-Progress
|
|
179
198
|
|
|
@@ -256,7 +275,7 @@ echo "You can now execute dependent tasks or continue with the full blueprint ex
|
|
|
256
275
|
|
|
257
276
|
## Error Handling
|
|
258
277
|
|
|
259
|
-
Read and execute
|
|
278
|
+
Read and execute $root/.ai/task-manager/config/hooks/POST_ERROR_DETECTION.md
|
|
260
279
|
|
|
261
280
|
## Usage Examples
|
|
262
281
|
|
|
@@ -2,20 +2,8 @@
|
|
|
2
2
|
argument-hint: "[testCommand]"
|
|
3
3
|
description: Fix the tests your task execution broke.
|
|
4
4
|
---
|
|
5
|
-
## Assistant Configuration
|
|
6
5
|
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
Load the following configuration files in order of precedence (later files override earlier ones):
|
|
10
|
-
1. `/workspace/AGENTS.md` - Project-level task management guidance
|
|
11
|
-
2. `/workspace/CLAUDE.md` - Claude-specific assistant configuration (if it exists)
|
|
12
|
-
3. `/home/node/.claude/CLAUDE.md` - Global Claude configuration from your home directory (if it exists)
|
|
13
|
-
|
|
14
|
-
These files contain your global and project-level configuration rules. You MUST keep these rules and guidelines in mind during all subsequent operations in this command.
|
|
15
|
-
|
|
16
|
-
---
|
|
17
|
-
|
|
18
|
-
Fix all failing tests in this repository. Think harder and use tools.
|
|
6
|
+
Fix all failing tests in this repository. Ultrathink, think harder, and use tools.
|
|
19
7
|
|
|
20
8
|
Execute this command to run the tests:
|
|
21
9
|
|
|
@@ -2,22 +2,53 @@
|
|
|
2
2
|
argument-hint: "[userPrompt]"
|
|
3
3
|
description: Execute the full workflow from plan creation to blueprint execution.
|
|
4
4
|
---
|
|
5
|
+
|
|
5
6
|
# Full Workflow Execution
|
|
6
7
|
|
|
7
8
|
You are a workflow composition assistant. Your role is to execute the complete task management workflow from plan creation through blueprint execution **without pausing between steps**. This is a fully automated workflow that executes all three steps sequentially.
|
|
8
9
|
|
|
9
|
-
|
|
10
|
+
---
|
|
10
11
|
|
|
11
|
-
|
|
12
|
+
## Find the AI Task Manager root
|
|
12
13
|
|
|
13
|
-
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
14
|
+
```bash
|
|
15
|
+
if [ ! -f /tmp/find-ai-task-manager-root.js ]; then
|
|
16
|
+
cat << 'EOF' > /tmp/find-ai-task-manager-root.js
|
|
17
|
+
const fs = require('fs');
|
|
18
|
+
const path = require('path');
|
|
19
|
+
|
|
20
|
+
const findRoot = (currentDir) => {
|
|
21
|
+
const taskManagerPath = path.join(currentDir, '.ai/task-manager');
|
|
22
|
+
const metadataPath = path.join(taskManagerPath, '.init-metadata.json');
|
|
23
|
+
|
|
24
|
+
try {
|
|
25
|
+
if (fs.existsSync(metadataPath) && JSON.parse(fs.readFileSync(metadataPath, 'utf8')).version) {
|
|
26
|
+
console.log(path.resolve(taskManagerPath));
|
|
27
|
+
process.exit(0);
|
|
28
|
+
}
|
|
29
|
+
} catch (e) {
|
|
30
|
+
// Continue searching
|
|
31
|
+
}
|
|
32
|
+
|
|
33
|
+
const parentDir = path.dirname(currentDir);
|
|
34
|
+
if (parentDir !== currentDir) {
|
|
35
|
+
findRoot(parentDir);
|
|
36
|
+
} else {
|
|
37
|
+
process.exit(1);
|
|
38
|
+
}
|
|
39
|
+
};
|
|
40
|
+
|
|
41
|
+
findRoot(process.cwd());
|
|
42
|
+
EOF
|
|
43
|
+
fi
|
|
17
44
|
|
|
18
|
-
|
|
45
|
+
root=$(node /tmp/find-ai-task-manager-root.js)
|
|
19
46
|
|
|
20
|
-
|
|
47
|
+
if [ -z "$root" ]; then
|
|
48
|
+
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
49
|
+
exit 1
|
|
50
|
+
fi
|
|
51
|
+
```
|
|
21
52
|
|
|
22
53
|
## Workflow Execution Instructions
|
|
23
54
|
|
|
@@ -52,24 +83,24 @@ Use your internal Todo task tool to track the workflow execution:
|
|
|
52
83
|
|
|
53
84
|
Execute the following plan creation process:
|
|
54
85
|
|
|
55
|
-
|
|
86
|
+
Ultrathink, think harder, and use tools.
|
|
56
87
|
|
|
57
88
|
You are a comprehensive task planning assistant. Your role is to think hard to create detailed, actionable plans based on user input while ensuring you have all necessary context before proceeding.
|
|
58
89
|
|
|
59
|
-
Include
|
|
90
|
+
Include $root/.ai/task-manager/config/TASK_MANAGER.md for the directory structure of tasks.
|
|
60
91
|
|
|
61
92
|
### Process
|
|
62
93
|
|
|
63
94
|
Use your internal Todo task tool to track the plan generation:
|
|
64
95
|
|
|
65
|
-
- [ ] Read and execute
|
|
96
|
+
- [ ] Read and execute $root/.ai/task-manager/config/hooks/PRE_PLAN.md
|
|
66
97
|
- [ ] User input and context analysis
|
|
67
98
|
- [ ] Clarification questions
|
|
68
99
|
- [ ] Plan generation: Executive Summary
|
|
69
100
|
- [ ] Plan generation: Detailed Steps
|
|
70
101
|
- [ ] Plan generation: Risk Considerations
|
|
71
102
|
- [ ] Plan generation: Success Metrics
|
|
72
|
-
- [ ] Read and execute
|
|
103
|
+
- [ ] Read and execute $root/.ai/task-manager/config/hooks/POST_PLAN.md
|
|
73
104
|
|
|
74
105
|
#### Context Analysis
|
|
75
106
|
Before creating any plan, analyze the user's request for:
|
|
@@ -125,7 +156,7 @@ This structured output enables automated workflow coordination and must be inclu
|
|
|
125
156
|
|
|
126
157
|
#### Plan Template
|
|
127
158
|
|
|
128
|
-
Use the template in
|
|
159
|
+
Use the template in $root/.ai/task-manager/config/templates/PLAN_TEMPLATE.md
|
|
129
160
|
|
|
130
161
|
#### Patterns to Avoid
|
|
131
162
|
Do not include the following in your plan output.
|
|
@@ -145,18 +176,8 @@ created: 2025-09-01
|
|
|
145
176
|
|
|
146
177
|
#### Plan ID Generation
|
|
147
178
|
|
|
148
|
-
|
|
179
|
+
Execute to auto-generate the next plan ID:
|
|
149
180
|
|
|
150
|
-
```bash
|
|
151
|
-
root=$(node -e 'const fs=require("fs"),path=require("path");const f=p=>{const t=path.join(p,".ai/task-manager");const m=path.join(t,".init-metadata.json");try{if(JSON.parse(fs.readFileSync(m)).version){console.log(path.resolve(t));process.exit(0)}}catch(e){};const d=path.dirname(p);if(d!==p)f(d)};f(process.cwd());process.exit(1)')
|
|
152
|
-
|
|
153
|
-
if [ -z "$root" ]; then
|
|
154
|
-
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
155
|
-
exit 1
|
|
156
|
-
fi
|
|
157
|
-
```
|
|
158
|
-
|
|
159
|
-
Then auto-generate the next plan ID:
|
|
160
181
|
```bash
|
|
161
182
|
node $root/config/scripts/get-next-plan-id.cjs
|
|
162
183
|
```
|
|
@@ -179,8 +200,6 @@ node $root/config/scripts/get-next-plan-id.cjs
|
|
|
179
200
|
|
|
180
201
|
Using the Plan ID extracted from Step 1, execute task generation:
|
|
181
202
|
|
|
182
|
-
Think harder and use tools.
|
|
183
|
-
|
|
184
203
|
You are a comprehensive task planning assistant. Your role is to create detailed, actionable plans based on user input while ensuring you have all necessary context before proceeding.
|
|
185
204
|
|
|
186
205
|
Include /TASK_MANAGER.md for the directory structure of tasks.
|
|
@@ -193,10 +212,10 @@ Use your internal Todo task tool to track the following process:
|
|
|
193
212
|
|
|
194
213
|
- [ ] Read and process plan [PLAN_ID from Step 1]
|
|
195
214
|
- [ ] Use the Task Generation Process to create tasks according to the Task Creation Guidelines
|
|
196
|
-
- [ ] Read and run the
|
|
215
|
+
- [ ] Read and run the $root/.ai/task-manager/config/hooks/POST_TASK_GENERATION_ALL.md
|
|
197
216
|
|
|
198
217
|
#### Input
|
|
199
|
-
- A plan document. See
|
|
218
|
+
- A plan document. See $root/.ai/task-manager/config/TASK_MANAGER.md to find the plan with ID from Step 1
|
|
200
219
|
- The plan contains high-level objectives and implementation steps
|
|
201
220
|
|
|
202
221
|
#### Input Error Handling
|
|
@@ -276,7 +295,7 @@ For each task, identify:
|
|
|
276
295
|
|
|
277
296
|
#### Step 3: Task Generation
|
|
278
297
|
|
|
279
|
-
Use the task template in
|
|
298
|
+
Use the task template in $root/.ai/task-manager/config/templates/TASK_TEMPLATE.md
|
|
280
299
|
|
|
281
300
|
**Task ID Generation:**
|
|
282
301
|
|
|
@@ -288,14 +307,14 @@ node $root/config/scripts/get-next-task-id.cjs [PLAN_ID from Step 1]
|
|
|
288
307
|
|
|
289
308
|
#### Step 4: POST_TASK_GENERATION_ALL hook
|
|
290
309
|
|
|
291
|
-
Read and run the
|
|
310
|
+
Read and run the $root/.ai/task-manager/config/hooks/POST_TASK_GENERATION_ALL.md
|
|
292
311
|
|
|
293
312
|
### Output Requirements
|
|
294
313
|
|
|
295
314
|
**Output Behavior:**
|
|
296
315
|
|
|
297
316
|
Provide a concise completion message with task count and location:
|
|
298
|
-
- Example: "Task generation complete. Created [count] tasks in
|
|
317
|
+
- Example: "Task generation complete. Created [count] tasks in `$root/.ai/task-manager/plans/[plan-id]--[name]/tasks/`"
|
|
299
318
|
|
|
300
319
|
**CRITICAL - Structured Output for Command Coordination:**
|
|
301
320
|
|
|
@@ -367,7 +386,7 @@ Use your internal Todo task tool to track the execution of all phases, and the f
|
|
|
367
386
|
|
|
368
387
|
#### Phase Pre-Execution
|
|
369
388
|
|
|
370
|
-
Read and execute
|
|
389
|
+
Read and execute $root/.ai/task-manager/config/hooks/PRE_PHASE.md
|
|
371
390
|
|
|
372
391
|
#### Phase Execution Workflow
|
|
373
392
|
|
|
@@ -376,7 +395,7 @@ Read and execute .ai/task-manager/config/hooks/PRE_PHASE.md
|
|
|
376
395
|
- List all tasks scheduled for parallel execution in this phase
|
|
377
396
|
|
|
378
397
|
2. **Agent Selection and Task Assignment**
|
|
379
|
-
Read and execute
|
|
398
|
+
Read and execute $root/.ai/task-manager/config/hooks/PRE_TASK_ASSIGNMENT.md
|
|
380
399
|
|
|
381
400
|
3. **Parallel Execution**
|
|
382
401
|
- Deploy all selected agents simultaneously using your internal Task tool
|
|
@@ -391,7 +410,7 @@ Read and execute .ai/task-manager/config/hooks/PRE_TASK_ASSIGNMENT.md
|
|
|
391
410
|
|
|
392
411
|
#### Phase Post-Execution
|
|
393
412
|
|
|
394
|
-
Read and execute
|
|
413
|
+
Read and execute $root/.ai/task-manager/config/hooks/POST_PHASE.md
|
|
395
414
|
|
|
396
415
|
#### Phase Transition
|
|
397
416
|
|
|
@@ -404,7 +423,7 @@ Read and execute .ai/task-manager/config/hooks/POST_PHASE.md
|
|
|
404
423
|
**Output Behavior:**
|
|
405
424
|
|
|
406
425
|
Provide a concise execution summary:
|
|
407
|
-
- Example: "Execution completed. Review summary:
|
|
426
|
+
- Example: "Execution completed. Review summary: `$root/.ai/task-manager/archive/[plan]/plan-[id].md`"
|
|
408
427
|
|
|
409
428
|
**CRITICAL - Structured Output for Command Coordination:**
|
|
410
429
|
|
|
@@ -415,7 +434,7 @@ Always end your output with a standardized summary in this exact format:
|
|
|
415
434
|
Execution Summary:
|
|
416
435
|
- Plan ID: [numeric-id]
|
|
417
436
|
- Status: Archived
|
|
418
|
-
- Location:
|
|
437
|
+
- Location: $root/.ai/task-manager/archive/[plan-id]--[plan-name]/
|
|
419
438
|
```
|
|
420
439
|
|
|
421
440
|
This structured output enables automated workflow coordination and must be included even when running standalone.
|
|
@@ -429,7 +448,7 @@ Upon successful completion of all phases and validation gates, perform the follo
|
|
|
429
448
|
|
|
430
449
|
#### Execution Summary Generation
|
|
431
450
|
|
|
432
|
-
Append an execution summary section to the plan document with the format described in
|
|
451
|
+
Append an execution summary section to the plan document with the format described in $root/.ai/task-manager/config/templates/EXECUTION_SUMMARY_TEMPLATE.md
|
|
433
452
|
|
|
434
453
|
#### Plan Archival
|
|
435
454
|
|
|
@@ -437,7 +456,7 @@ After successfully appending the execution summary:
|
|
|
437
456
|
|
|
438
457
|
**Move completed plan to archive**:
|
|
439
458
|
```bash
|
|
440
|
-
mv
|
|
459
|
+
mv $root/.ai/task-manager/plans/[plan-folder] $root/.ai/task-manager/archive/
|
|
441
460
|
```
|
|
442
461
|
|
|
443
462
|
**Progress**: ⬛⬛⬛ 100% - Step 3/3: Blueprint Execution Complete
|
|
@@ -5,24 +5,54 @@ description: Generate tasks to implement the plan with the provided ID.
|
|
|
5
5
|
|
|
6
6
|
# Comprehensive Task List Creation
|
|
7
7
|
|
|
8
|
-
|
|
8
|
+
---
|
|
9
9
|
|
|
10
|
-
|
|
10
|
+
Ultrathink, think harder, and use tools.
|
|
11
11
|
|
|
12
|
-
|
|
13
|
-
1. `/workspace/AGENTS.md` - Project-level task management guidance
|
|
14
|
-
2. `/workspace/CLAUDE.md` - Claude-specific assistant configuration (if it exists)
|
|
15
|
-
3. `/home/node/.claude/CLAUDE.md` - Global Claude configuration from your home directory (if it exists)
|
|
12
|
+
## Find the AI Task Manager root
|
|
16
13
|
|
|
17
|
-
|
|
14
|
+
```bash
|
|
15
|
+
if [ ! -f /tmp/find-ai-task-manager-root.js ]; then
|
|
16
|
+
cat << 'EOF' > /tmp/find-ai-task-manager-root.js
|
|
17
|
+
const fs = require('fs');
|
|
18
|
+
const path = require('path');
|
|
19
|
+
|
|
20
|
+
const findRoot = (currentDir) => {
|
|
21
|
+
const taskManagerPath = path.join(currentDir, '.ai/task-manager');
|
|
22
|
+
const metadataPath = path.join(taskManagerPath, '.init-metadata.json');
|
|
23
|
+
|
|
24
|
+
try {
|
|
25
|
+
if (fs.existsSync(metadataPath) && JSON.parse(fs.readFileSync(metadataPath, 'utf8')).version) {
|
|
26
|
+
console.log(path.resolve(taskManagerPath));
|
|
27
|
+
process.exit(0);
|
|
28
|
+
}
|
|
29
|
+
} catch (e) {
|
|
30
|
+
// Continue searching
|
|
31
|
+
}
|
|
32
|
+
|
|
33
|
+
const parentDir = path.dirname(currentDir);
|
|
34
|
+
if (parentDir !== currentDir) {
|
|
35
|
+
findRoot(parentDir);
|
|
36
|
+
} else {
|
|
37
|
+
process.exit(1);
|
|
38
|
+
}
|
|
39
|
+
};
|
|
40
|
+
|
|
41
|
+
findRoot(process.cwd());
|
|
42
|
+
EOF
|
|
43
|
+
fi
|
|
18
44
|
|
|
19
|
-
|
|
45
|
+
root=$(node /tmp/find-ai-task-manager-root.js)
|
|
20
46
|
|
|
21
|
-
|
|
47
|
+
if [ -z "$root" ]; then
|
|
48
|
+
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
49
|
+
exit 1
|
|
50
|
+
fi
|
|
51
|
+
```
|
|
22
52
|
|
|
23
53
|
You are a comprehensive task planning assistant. Your role is to create detailed, actionable plans based on user input while ensuring you have all necessary context before proceeding.
|
|
24
54
|
|
|
25
|
-
Include
|
|
55
|
+
Include `$root/.ai/task-manager/config/TASK_MANAGER.md` for the directory structure of tasks.
|
|
26
56
|
|
|
27
57
|
## Instructions
|
|
28
58
|
|
|
@@ -32,25 +62,12 @@ Use your internal Todo task tool to track the following process:
|
|
|
32
62
|
|
|
33
63
|
- [ ] Read and process plan $1
|
|
34
64
|
- [ ] Use the Task Generation Process to create tasks according to the Task Creation Guidelines
|
|
35
|
-
- [ ] Read and run the
|
|
65
|
+
- [ ] Read and run the $root/.ai/task-manager/config/hooks/POST_TASK_GENERATION_ALL.md
|
|
36
66
|
|
|
37
67
|
### Input
|
|
38
68
|
|
|
39
69
|
- A plan document. Extract it with the following command.
|
|
40
70
|
|
|
41
|
-
First, discover the task manager root directory:
|
|
42
|
-
|
|
43
|
-
```bash
|
|
44
|
-
root=$(node -e 'const fs=require("fs"),path=require("path");const f=p=>{const t=path.join(p,".ai/task-manager");const m=path.join(t,".init-metadata.json");try{if(JSON.parse(fs.readFileSync(m)).version){console.log(path.resolve(t));process.exit(0)}}catch(e){};const d=path.dirname(p);if(d!==p)f(d)};f(process.cwd());process.exit(1)')
|
|
45
|
-
|
|
46
|
-
if [ -z "$root" ]; then
|
|
47
|
-
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
48
|
-
exit 1
|
|
49
|
-
fi
|
|
50
|
-
```
|
|
51
|
-
|
|
52
|
-
Then extract validation results:
|
|
53
|
-
|
|
54
71
|
```bash
|
|
55
72
|
# Extract validation results directly from script
|
|
56
73
|
plan_file=$(node $root/config/scripts/validate-plan-blueprint.cjs $1 planFile)
|
|
@@ -249,7 +266,7 @@ The schema for this frontmatter is:
|
|
|
249
266
|
|
|
250
267
|
##### Task Body Structure
|
|
251
268
|
|
|
252
|
-
Use the task template in
|
|
269
|
+
Use the task template in $root/.ai/task-manager/config/templates/TASK_TEMPLATE.md
|
|
253
270
|
|
|
254
271
|
##### Task ID Generation
|
|
255
272
|
|
|
@@ -309,14 +326,14 @@ If the plan lacks sufficient detail:
|
|
|
309
326
|
|
|
310
327
|
#### Step 4: POST_TASK_GENERATION_ALL hook
|
|
311
328
|
|
|
312
|
-
Read and run the
|
|
329
|
+
Read and run the $root/.ai/task-manager/config/hooks/POST_TASK_GENERATION_ALL.md
|
|
313
330
|
|
|
314
331
|
### Output Requirements
|
|
315
332
|
|
|
316
333
|
**Output Behavior:**
|
|
317
334
|
|
|
318
335
|
Provide a concise completion message with task count and location:
|
|
319
|
-
- Example: "Task generation complete. Created [count] tasks in
|
|
336
|
+
- Example: "Task generation complete. Created [count] tasks in `$root/.ai/task-manager/plans/[plan-id]--[name]/tasks/`"
|
|
320
337
|
|
|
321
338
|
**CRITICAL - Structured Output for Command Coordination:**
|
|
322
339
|
|
|
@@ -9,22 +9,52 @@ refining the document so that task generators receive the clearest possible inst
|
|
|
9
9
|
work product of another assistant: your responsibility is to pressure test it, request any missing information from the
|
|
10
10
|
user, and update the plan with the refinements. Use the plan-creator sub-agent for this if it is available.
|
|
11
11
|
|
|
12
|
-
|
|
12
|
+
---
|
|
13
13
|
|
|
14
|
-
|
|
14
|
+
Ultrathink, think harder, and use tools.
|
|
15
15
|
|
|
16
|
-
|
|
17
|
-
1. `/workspace/AGENTS.md` - Project-level task management guidance
|
|
18
|
-
2. `/workspace/CLAUDE.md` - Claude-specific assistant configuration (if it exists)
|
|
19
|
-
3. `/home/node/.claude/CLAUDE.md` - Global Claude configuration from your home directory (if it exists)
|
|
16
|
+
## Find the AI Task Manager root
|
|
20
17
|
|
|
21
|
-
|
|
18
|
+
```bash
|
|
19
|
+
if [ ! -f /tmp/find-ai-task-manager-root.js ]; then
|
|
20
|
+
cat << 'EOF' > /tmp/find-ai-task-manager-root.js
|
|
21
|
+
const fs = require('fs');
|
|
22
|
+
const path = require('path');
|
|
23
|
+
|
|
24
|
+
const findRoot = (currentDir) => {
|
|
25
|
+
const taskManagerPath = path.join(currentDir, '.ai/task-manager');
|
|
26
|
+
const metadataPath = path.join(taskManagerPath, '.init-metadata.json');
|
|
27
|
+
|
|
28
|
+
try {
|
|
29
|
+
if (fs.existsSync(metadataPath) && JSON.parse(fs.readFileSync(metadataPath, 'utf8')).version) {
|
|
30
|
+
console.log(path.resolve(taskManagerPath));
|
|
31
|
+
process.exit(0);
|
|
32
|
+
}
|
|
33
|
+
} catch (e) {
|
|
34
|
+
// Continue searching
|
|
35
|
+
}
|
|
36
|
+
|
|
37
|
+
const parentDir = path.dirname(currentDir);
|
|
38
|
+
if (parentDir !== currentDir) {
|
|
39
|
+
findRoot(parentDir);
|
|
40
|
+
} else {
|
|
41
|
+
process.exit(1);
|
|
42
|
+
}
|
|
43
|
+
};
|
|
44
|
+
|
|
45
|
+
findRoot(process.cwd());
|
|
46
|
+
EOF
|
|
47
|
+
fi
|
|
22
48
|
|
|
23
|
-
|
|
49
|
+
root=$(node /tmp/find-ai-task-manager-root.js)
|
|
24
50
|
|
|
25
|
-
|
|
51
|
+
if [ -z "$root" ]; then
|
|
52
|
+
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
53
|
+
exit 1
|
|
54
|
+
fi
|
|
55
|
+
```
|
|
26
56
|
|
|
27
|
-
Include
|
|
57
|
+
Include `$root/.ai/task-manager/config/TASK_MANAGER.md` to understand the plan directory structure and naming conventions.
|
|
28
58
|
|
|
29
59
|
## Inputs
|
|
30
60
|
- **Plan ID**: `$1` (required)
|
|
@@ -34,18 +64,7 @@ If the plan ID is missing, immediately stop and show an error explaining correct
|
|
|
34
64
|
|
|
35
65
|
### Plan Discovery and Validation
|
|
36
66
|
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
```bash
|
|
40
|
-
root=$(node -e 'const fs=require("fs"),path=require("path");const f=p=>{const t=path.join(p,".ai/task-manager");const m=path.join(t,".init-metadata.json");try{if(JSON.parse(fs.readFileSync(m)).version){console.log(path.resolve(t));process.exit(0)}}catch(e){};const d=path.dirname(p);if(d!==p)f(d)};f(process.cwd());process.exit(1)')
|
|
41
|
-
|
|
42
|
-
if [ -z "$root" ]; then
|
|
43
|
-
echo "Error: Could not find task manager root directory (.ai/task-manager)"
|
|
44
|
-
exit 1
|
|
45
|
-
fi
|
|
46
|
-
```
|
|
47
|
-
|
|
48
|
-
Then obtain the plan using the plan ID:
|
|
67
|
+
Obtain the plan using the plan ID:
|
|
49
68
|
|
|
50
69
|
```bash
|
|
51
70
|
# Extract validation results directly from script
|
|
@@ -56,17 +75,17 @@ plan_file=$(node $root/config/scripts/validate-plan-blueprint.cjs $1 planFile)
|
|
|
56
75
|
|
|
57
76
|
Use your internal Todo tool to track the entire refinement workflow:
|
|
58
77
|
|
|
59
|
-
- [ ] Load
|
|
78
|
+
- [ ] Load `$root/.ai/task-manager/config/hooks/PRE_PLAN.md`
|
|
60
79
|
- [ ] Stage 1: Baseline Review
|
|
61
80
|
- [ ] Stage 2: Clarification Loop
|
|
62
81
|
- [ ] Stage 3: Refinement Implementation
|
|
63
82
|
- [ ] Review the existing plan end-to-end (frontmatter, clarifications, architecture, risks, etc.)
|
|
64
83
|
- [ ] Surface strengths, contradictions, and potential risks, without updating the plan
|
|
65
84
|
- [ ] Use the "Total Clarification Algorithm" to get the missing clarification from the user
|
|
66
|
-
- [ ] Apply refinements using
|
|
85
|
+
- [ ] Apply refinements using `$root/.ai/task-manager/config/templates/PLAN_TEMPLATE.md` as the structural baseline
|
|
67
86
|
- [ ] Update the "Plan Clarifications" table with the latest Q&A
|
|
68
87
|
- [ ] Update the plan file (stored in `plan_file`) with the refinements from steps above
|
|
69
|
-
- [ ] Re-run
|
|
88
|
+
- [ ] Re-run `$root/.ai/task-manager/config/hooks/POST_PLAN.md`
|
|
70
89
|
|
|
71
90
|
## Stage 1: Baseline Review
|
|
72
91
|
|
|
@@ -109,7 +128,7 @@ flowchart TD
|
|
|
109
128
|
Once you have sufficient context (or have documented the missing context), refine the plan directly in-place:
|
|
110
129
|
|
|
111
130
|
1. **Maintain Identity**: Keep the existing `id` and directory. Do not create a new plan ID.
|
|
112
|
-
2. **Structure Compliance**: Ensure the plan still follows
|
|
131
|
+
2. **Structure Compliance**: Ensure the plan still follows `$root/.ai/task-manager/config/templates/PLAN_TEMPLATE.md`. Add missing sections if necessary.
|
|
113
132
|
3. **Content Updates**:
|
|
114
133
|
- Refresh the executive summary to reflect clarifications and new insights.
|
|
115
134
|
- Update architectural sections, diagrams, and risk mitigations to resolve the identified gaps.
|
|
@@ -117,7 +136,7 @@ Once you have sufficient context (or have documented the missing context), refin
|
|
|
117
136
|
- Clearly reference clarifications in the relevant plan sections (e.g., italicized notes that point back to the Q&A table).
|
|
118
137
|
4. **Net-New Sections**: If the plan needs a new subsection (e.g., Decision Log, Data Contracts), add it under `Notes` with a clearly labeled section so it remains discoverable.
|
|
119
138
|
5. **Change Log**: Append a bullet list in the `Notes` section that briefly states what changed in this refinement session (e.g., `- 2025-03-16: Clarified auth flow tokens and updated architecture diagram`).
|
|
120
|
-
6. **Validation Hooks**: Execute
|
|
139
|
+
6. **Validation Hooks**: Execute `$root/.ai/task-manager/config/hooks/POST_PLAN.md` to ensure the refined plan still meets quality bars.
|
|
121
140
|
|
|
122
141
|
## Output Requirements
|
|
123
142
|
|