thanh-kit 2.5.1 → 2.5.2

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 (123) hide show
  1. package/dist/index.js +20 -61
  2. package/dist/index.js.map +1 -1
  3. package/package.json +1 -1
  4. package/templates/{commands → command-archive}/ask.md +5 -5
  5. package/templates/{commands → command-archive}/ck-help.md +18 -2
  6. package/templates/command-archive/docs/init.md +38 -0
  7. package/templates/command-archive/docs/summarize.md +22 -0
  8. package/templates/command-archive/docs/update.md +76 -0
  9. package/templates/command-archive/journal.md +18 -0
  10. package/templates/{commands → command-archive}/kanban.md +5 -7
  11. package/templates/{commands → command-archive}/plan/archive.md +2 -2
  12. package/templates/command-archive/plan/red-team.md +200 -0
  13. package/templates/command-archive/plan/validate.md +188 -0
  14. package/templates/command-archive/preview.md +283 -0
  15. package/templates/command-archive/review/codebase/parallel.md +122 -0
  16. package/templates/{commands → command-archive}/test/ui.md +3 -3
  17. package/templates/{commands → command-archive}/use-mcp.md +6 -2
  18. package/templates/command-archive/worktree.md +109 -0
  19. package/templates/{workflows → rules}/development-rules.md +12 -53
  20. package/templates/rules/orchestration-protocol.md +43 -0
  21. package/templates/{workflows → rules}/primary-workflow.md +16 -4
  22. package/templates/rules/team-coordination-rules.md +90 -0
  23. package/templates/schemas/ck-config.schema.json +381 -0
  24. package/templates/commands/README.md +0 -251
  25. package/templates/commands/bootstrap/auto/fast.md +0 -111
  26. package/templates/commands/bootstrap/auto/parallel.md +0 -66
  27. package/templates/commands/bootstrap/auto.md +0 -115
  28. package/templates/commands/bootstrap.md +0 -137
  29. package/templates/commands/brainstorm.md +0 -74
  30. package/templates/commands/build.md +0 -39
  31. package/templates/commands/checkpoint.md +0 -156
  32. package/templates/commands/code/auto.md +0 -170
  33. package/templates/commands/code/no-test.md +0 -158
  34. package/templates/commands/code/parallel.md +0 -55
  35. package/templates/commands/code-simplifier.md +0 -71
  36. package/templates/commands/code.md +0 -176
  37. package/templates/commands/compact.md +0 -57
  38. package/templates/commands/content/cro.md +0 -43
  39. package/templates/commands/content/enhance.md +0 -14
  40. package/templates/commands/content/fast.md +0 -13
  41. package/templates/commands/content/good.md +0 -16
  42. package/templates/commands/context.md +0 -48
  43. package/templates/commands/cook/auto/fast.md +0 -26
  44. package/templates/commands/cook/auto/parallel.md +0 -49
  45. package/templates/commands/cook/auto.md +0 -15
  46. package/templates/commands/cook/fast.md +0 -47
  47. package/templates/commands/cook/hard.md +0 -80
  48. package/templates/commands/cook/parallel.md +0 -90
  49. package/templates/commands/cook.md +0 -105
  50. package/templates/commands/create-feature.md +0 -48
  51. package/templates/commands/db-migrate.md +0 -52
  52. package/templates/commands/debug.md +0 -13
  53. package/templates/commands/design/3d.md +0 -83
  54. package/templates/commands/design/describe.md +0 -23
  55. package/templates/commands/design/fast.md +0 -31
  56. package/templates/commands/design/good.md +0 -35
  57. package/templates/commands/design/screenshot.md +0 -34
  58. package/templates/commands/design/video.md +0 -34
  59. package/templates/commands/docs/init.md +0 -39
  60. package/templates/commands/docs/summarize.md +0 -31
  61. package/templates/commands/docs/update.md +0 -57
  62. package/templates/commands/feature.md +0 -62
  63. package/templates/commands/fix/ci.md +0 -17
  64. package/templates/commands/fix/fast.md +0 -19
  65. package/templates/commands/fix/hard.md +0 -39
  66. package/templates/commands/fix/logs.md +0 -26
  67. package/templates/commands/fix/parallel.md +0 -54
  68. package/templates/commands/fix/test.md +0 -20
  69. package/templates/commands/fix/types.md +0 -9
  70. package/templates/commands/fix/ui.md +0 -48
  71. package/templates/commands/fix-issue.md +0 -177
  72. package/templates/commands/fix.md +0 -43
  73. package/templates/commands/generate-dto.md +0 -67
  74. package/templates/commands/git/cm.md +0 -5
  75. package/templates/commands/git/cp.md +0 -4
  76. package/templates/commands/git/merge.md +0 -40
  77. package/templates/commands/git/pr.md +0 -48
  78. package/templates/commands/integrate/polar.md +0 -28
  79. package/templates/commands/integrate/sepay.md +0 -28
  80. package/templates/commands/investigate.md +0 -324
  81. package/templates/commands/journal.md +0 -7
  82. package/templates/commands/lint.md +0 -47
  83. package/templates/commands/migration.md +0 -111
  84. package/templates/commands/performance.md +0 -110
  85. package/templates/commands/plan/ci.md +0 -33
  86. package/templates/commands/plan/cro.md +0 -69
  87. package/templates/commands/plan/fast.md +0 -86
  88. package/templates/commands/plan/hard.md +0 -103
  89. package/templates/commands/plan/parallel.md +0 -152
  90. package/templates/commands/plan/preview.md +0 -40
  91. package/templates/commands/plan/two.md +0 -52
  92. package/templates/commands/plan/validate.md +0 -132
  93. package/templates/commands/plan.md +0 -36
  94. package/templates/commands/pr.md +0 -49
  95. package/templates/commands/preview.md +0 -87
  96. package/templates/commands/release-notes.md +0 -144
  97. package/templates/commands/review/post-task.md +0 -157
  98. package/templates/commands/review-changes.md +0 -46
  99. package/templates/commands/review.md +0 -56
  100. package/templates/commands/scout/ext.md +0 -35
  101. package/templates/commands/scout.md +0 -283
  102. package/templates/commands/security.md +0 -119
  103. package/templates/commands/skill/add.md +0 -36
  104. package/templates/commands/skill/create.md +0 -29
  105. package/templates/commands/skill/fix-logs.md +0 -22
  106. package/templates/commands/skill/optimize/auto.md +0 -25
  107. package/templates/commands/skill/optimize.md +0 -34
  108. package/templates/commands/skill/plan.md +0 -45
  109. package/templates/commands/worktree.md +0 -126
  110. package/templates/memory/session-log.md +0 -186
  111. package/templates/router/README.md +0 -294
  112. package/templates/router/agents-guide.md +0 -38
  113. package/templates/router/commands-guide.md +0 -122
  114. package/templates/router/decision-flow.md +0 -92
  115. package/templates/router/skills-guide.md +0 -127
  116. package/templates/router/workflows-guide.md +0 -68
  117. package/templates/workflows/README.md +0 -241
  118. package/templates/workflows/orchestration-protocol.md +0 -16
  119. /package/templates/{commands → command-archive}/coding-level.md +0 -0
  120. /package/templates/{commands → command-archive}/review/codebase.md +0 -0
  121. /package/templates/{commands → command-archive}/test.md +0 -0
  122. /package/templates/{commands → command-archive}/watzup.md +0 -0
  123. /package/templates/{workflows → rules}/documentation-management.md +0 -0
@@ -1,87 +0,0 @@
1
- ---
2
- description: Path to markdown file, plan directory, or plans collection
3
- arguments:
4
- - name: path
5
- description: Path to file or directory to preview
6
- required: false
7
- ---
8
-
9
- Universal viewer using `markdown-novel-viewer` skill - pass ANY path and see it rendered nicely.
10
-
11
- ## Usage
12
-
13
- - `/preview <file.md>` - View markdown file in novel-reader UI
14
- - `/preview <directory/>` - Browse directory contents
15
- - `/preview --stop` - Stop running server
16
-
17
- ## Examples
18
-
19
- ```bash
20
- /preview plans/my-plan/plan.md # View markdown file
21
- /preview plans/ # Browse plans directory
22
- /preview docs/ # Browse docs directory
23
- /preview any/path/to/file.md # View any markdown file
24
- /preview any/path/ # Browse any directory
25
- ```
26
-
27
- ## Execution
28
-
29
- **IMPORTANT:** Run server as Claude Code background task using `run_in_background: true` with the Bash tool. This makes the server visible in `/tasks` and manageable via `KillShell`.
30
-
31
- Check if this script is located in the current workspace or in `$HOME/.claude/skills/markdown-novel-viewer` directory:
32
- - If in current workspace: `$SKILL_DIR_PATH` = `./.claude/skills/markdown-novel-viewer/`
33
- - If in home directory: `$SKILL_DIR_PATH` = `$HOME/.claude/skills/markdown-novel-viewer/`
34
-
35
- ### Stop Server
36
-
37
- If `--stop` flag is provided:
38
-
39
- ```bash
40
- node $SKILL_DIR_PATH/scripts/server.cjs --stop
41
- ```
42
-
43
- ### Start Server
44
-
45
- Otherwise, run the `markdown-novel-viewer` server as CC background task with `--foreground` flag (keeps process alive for CC task management):
46
-
47
- ```bash
48
- # Determine if path is file or directory
49
- INPUT_PATH="{{path}}"
50
- if [[ -d "$INPUT_PATH" ]]; then
51
- # Directory mode - browse
52
- node $SKILL_DIR_PATH/scripts/server.cjs \
53
- --dir "$INPUT_PATH" \
54
- --host 0.0.0.0 \
55
- --open \
56
- --foreground
57
- else
58
- # File mode - view markdown
59
- node $SKILL_DIR_PATH/scripts/server.cjs \
60
- --file "$INPUT_PATH" \
61
- --host 0.0.0.0 \
62
- --open \
63
- --foreground
64
- fi
65
- ```
66
-
67
- **Critical:** When calling the Bash tool:
68
- - Set `run_in_background: true` to run as CC background task
69
- - Set `timeout: 300000` (5 minutes) to prevent premature termination
70
- - Parse JSON output and report URL to user
71
-
72
- Example Bash tool call:
73
- ```json
74
- {
75
- "command": "node .claude/skills/markdown-novel-viewer/scripts/server.cjs --dir \"path\" --host 0.0.0.0 --open --foreground",
76
- "run_in_background": true,
77
- "timeout": 300000,
78
- "description": "Start preview server in background"
79
- }
80
- ```
81
-
82
- After starting, parse the JSON output (e.g., `{"success":true,"url":"http://localhost:3456/view?file=...","networkUrl":"http://192.168.1.x:3456/view?file=..."}`) and report:
83
- - Local URL for browser access
84
- - Network URL for remote device access (if available)
85
- - Inform user that server is now running as CC background task (visible in `/tasks`)
86
-
87
- **CRITICAL:** MUST display the FULL URL including path and query string (e.g., `http://localhost:3456/view?file=/path/to/file.md`). NEVER truncate to just `host:port` (e.g., `http://localhost:3456`). The full URL is required for direct file access.
@@ -1,144 +0,0 @@
1
- # Generate Release Notes: $ARGUMENTS
2
-
3
- Generate or update release notes for a feature or PR.
4
-
5
- ## Phase 1: Determine Source
6
-
7
- Parse input from: $ARGUMENTS
8
-
9
- **Source Options:**
10
- 1. `--source=<path>` - Use feature documentation file
11
- 2. `--compare=<base>:<head>` - Compare git branches
12
- 3. `--investigate` - Manual investigation mode
13
-
14
- If no option specified, prompt for source.
15
-
16
- ## Phase 2: Gather Information
17
-
18
- ### Option A: From Feature Documentation
19
- ```bash
20
- # Read the feature documentation
21
- cat <source-path>
22
- ```
23
-
24
- Extract:
25
- - Feature name and description
26
- - Key capabilities
27
- - Technical implementation
28
- - API endpoints
29
- - UI components
30
-
31
- ### Option B: From Branch Comparison
32
- ```bash
33
- # Fetch latest
34
- git fetch origin
35
-
36
- # Get commits between branches
37
- git log origin/<base>...origin/<head> --oneline
38
-
39
- # Get file changes
40
- git diff origin/<base>...origin/<head> --stat
41
-
42
- # Get detailed diff
43
- git diff origin/<base>...origin/<head>
44
- ```
45
-
46
- Extract:
47
- - Changed files
48
- - New features
49
- - Bug fixes
50
- - Breaking changes
51
-
52
- ### Option C: Investigation Mode
53
- 1. Ask user for feature scope
54
- 2. Search codebase for relevant files
55
- 3. Analyze implementation
56
-
57
- ## Phase 3: Generate Release Note
58
-
59
- Create file at: `docs/release-notes/YYMMDD-<feature-slug>.md`
60
-
61
- **Filename Format:**
62
- - YY = 2-digit year
63
- - MM = 2-digit month
64
- - DD = 2-digit day
65
- - feature-slug = kebab-case feature name
66
-
67
- **Template:**
68
- ```markdown
69
- # Release Notes: [Feature Name]
70
-
71
- **Date:** YYYY-MM-DD
72
- **Version:** X.Y.Z (if applicable)
73
- **Status:** Draft | Released
74
-
75
- ---
76
-
77
- ## Summary
78
-
79
- [One paragraph describing the feature and its value to users]
80
-
81
- ## New Features
82
-
83
- - **[Feature 1]**: Brief description
84
- - **[Feature 2]**: Brief description
85
-
86
- ## Improvements
87
-
88
- - **[Improvement 1]**: Brief description
89
-
90
- ## Bug Fixes
91
-
92
- - **[Fix 1]**: Brief description
93
-
94
- ## Breaking Changes
95
-
96
- > **Warning**: Breaking changes require migration
97
-
98
- - [Breaking change description and migration path]
99
-
100
- ## Technical Details
101
-
102
- ### Backend
103
- - [Service changes]
104
- - [API changes]
105
-
106
- ### Frontend
107
- - [Component changes]
108
- - [UI changes]
109
-
110
- ### Database
111
- - [Migration info]
112
-
113
- ## Related Documentation
114
-
115
- - [Feature Documentation](../business-features/...)
116
- - [API Reference](../api/...)
117
-
118
- ## Contributors
119
-
120
- - @contributor
121
-
122
- ---
123
-
124
- *Generated with [Claude Code](https://claude.com/claude-code)*
125
- ```
126
-
127
- ## Phase 4: Review & Finalize
128
-
129
- 1. Present the generated release note
130
- 2. Ask for any corrections or additions
131
- 3. Save the file
132
-
133
- ## Examples
134
-
135
- ```bash
136
- # From feature docs
137
- /release-notes kudos --source=docs/business-features/TextSnippet/detailed-features/README.KudosFeature.md
138
-
139
- # From branch comparison
140
- /release-notes employee-export --compare=feature/employee-export:develop
141
-
142
- # Investigation mode
143
- /release-notes authentication --investigate
144
- ```
@@ -1,157 +0,0 @@
1
- ---
2
- description: ⚡⚡⚡ Two-pass code review for task completion
3
- argument-hint: [optional-focus-area]
4
- ---
5
-
6
- # Post-Task Two-Pass Code Review
7
-
8
- Execute mandatory two-pass review protocol after completing code changes.
9
- Focus: $ARGUMENTS
10
-
11
- ---
12
-
13
- ## Protocol Overview
14
-
15
- This review ensures ALL code changes follow best practices and project conventions.
16
- Two passes guarantee corrections don't introduce new issues.
17
-
18
- ---
19
-
20
- ## Pass 1: Initial Review
21
-
22
- ### Step 1.1: Gather Changes
23
- ```bash
24
- # Get unstaged changes summary
25
- git diff --stat
26
-
27
- # Get detailed changes
28
- git diff
29
- ```
30
-
31
- ### Step 1.2: Review Checklist
32
-
33
- **Task Correctness:**
34
- - [ ] Changes achieve the intended task objective
35
- - [ ] No unrelated/unnecessary modifications
36
- - [ ] Edge cases handled appropriately
37
- - [ ] Error handling is complete
38
-
39
- **Code Standards (EasyPlatform):**
40
-
41
- *Backend:*
42
- - [ ] Uses platform repository (IPlatformQueryableRootRepository)
43
- - [ ] Uses PlatformValidationResult fluent API (.And(), .AndAsync())
44
- - [ ] Side effects in Entity Event Handlers, not command handlers
45
- - [ ] DTOs own mapping via MapToEntity()/MapToObject()
46
- - [ ] Command + Result + Handler in ONE file
47
-
48
- *Frontend:*
49
- - [ ] Extends appropriate base class (AppBaseComponent, AppBaseVmStoreComponent, AppBaseFormComponent)
50
- - [ ] Uses PlatformVmStore for state management
51
- - [ ] Uses .pipe(this.untilDestroyed()) for subscriptions
52
- - [ ] All template elements have BEM classes
53
-
54
- **Architecture:**
55
- - [ ] Logic placed in lowest appropriate layer (Entity > Service > Component)
56
- - [ ] No cross-service direct database access
57
- - [ ] Follows existing patterns found in codebase
58
-
59
- **Security:**
60
- - [ ] No hardcoded secrets or credentials
61
- - [ ] Input validation at boundaries
62
- - [ ] Proper authorization checks
63
-
64
- ### Step 1.3: Execute Corrections
65
- If any issues found:
66
- 1. Fix each issue directly
67
- 2. Mark `PASS_1_MADE_CHANGES = true`
68
- 3. Proceed to Pass 2
69
-
70
- If no issues found:
71
- 1. Mark `PASS_1_MADE_CHANGES = false`
72
- 2. Skip to Final Report
73
-
74
- ---
75
-
76
- ## Pass 2: Re-Review (Conditional)
77
-
78
- **ONLY EXECUTE IF PASS 1 MADE CHANGES**
79
-
80
- ### Step 2.1: Verify All Changes
81
- ```bash
82
- # Check updated changes
83
- git diff --stat
84
- git diff
85
- ```
86
-
87
- ### Step 2.2: Full Re-Review
88
- Repeat Pass 1 checklist on ALL current changes (original + corrections).
89
-
90
- Focus areas:
91
- - [ ] Original task still correctly implemented
92
- - [ ] Corrections are valid and complete
93
- - [ ] No new issues introduced by corrections
94
- - [ ] Code is production-ready
95
-
96
- ### Step 2.3: Final Corrections
97
- If issues found:
98
- - Apply minimal, targeted fixes
99
- - Document any trade-offs made
100
-
101
- ---
102
-
103
- ## Final Report
104
-
105
- ### Summary Template
106
- ```markdown
107
- ## Post-Task Review Complete
108
-
109
- **Task:** [Brief description]
110
- **Pass 1 Result:** [Clean / N issues fixed]
111
- **Pass 2 Required:** [Yes/No]
112
- **Pass 2 Result:** [N/A / Clean / N issues fixed]
113
-
114
- ### Changes Summary
115
- - [List of files modified]
116
- - [Key changes made]
117
-
118
- ### Issues Found & Fixed
119
- 1. [Issue]: [Fix applied]
120
- 2. ...
121
-
122
- ### Remaining Concerns (if any)
123
- - [Concern]: [Reason not addressed / Recommended follow-up]
124
-
125
- ### Verification
126
- - [ ] Task objective achieved
127
- - [ ] Code follows project conventions
128
- - [ ] No security vulnerabilities
129
- - [ ] Ready for commit
130
- ```
131
-
132
- ---
133
-
134
- ## Quick Commands
135
-
136
- ```bash
137
- # View changes for review
138
- git diff --stat && git diff
139
-
140
- # Check specific file
141
- git diff path/to/file.ts
142
-
143
- # Discard a problematic change
144
- git checkout -- path/to/file.ts
145
-
146
- # Stage reviewed changes
147
- git add -p # Interactive staging
148
- ```
149
-
150
- ---
151
-
152
- ## Integration Notes
153
-
154
- - This command is auto-triggered by workflow orchestration after `/cook`, `/fix`, `/code`
155
- - Can be manually invoked anytime with `/review:post-task`
156
- - For PR reviews, use `/review:codebase` instead
157
- - Use `code-reviewer` subagent for complex reviews requiring deeper analysis
@@ -1,46 +0,0 @@
1
- ---
2
- description: Review all uncommitted changes before commit
3
- allowed-tools: Bash, Read, Glob, Grep, TodoWrite
4
- ---
5
-
6
- Review all uncommitted changes and provide feedback.
7
-
8
- ## Steps:
9
-
10
- 1. **Get Change Summary**
11
- - Run `git status` to see all changed files
12
- - Run `git diff` to see actual changes (staged and unstaged)
13
-
14
- 2. **Analyze Each Changed File**
15
- For each modified file, check:
16
- - Does it follow platform patterns from CLAUDE.md?
17
- - Any security concerns (hardcoded secrets, SQL injection, XSS)?
18
- - Any performance issues (N+1 queries, missing async/await)?
19
- - Proper error handling?
20
- - Code style compliance?
21
-
22
- 3. **Check for Common Issues**
23
- - Unused imports or variables
24
- - Console.log/Debug.WriteLine statements left in
25
- - Hardcoded values that should be configuration
26
- - Missing async/await keywords
27
- - Incorrect exception handling
28
- - Missing validation
29
-
30
- 4. **Backend-Specific Checks**
31
- - CQRS patterns followed correctly
32
- - Repository usage (no direct DbContext access)
33
- - Entity DTO mapping patterns
34
- - Validation using PlatformValidationResult
35
-
36
- 5. **Frontend-Specific Checks**
37
- - Component base class inheritance correct
38
- - State management patterns
39
- - Memory leaks (missing unsubscribe)
40
- - Template binding issues
41
-
42
- 6. **Provide Summary**
43
- - **Critical Issues**: Must fix before commit
44
- - **Warnings**: Should consider fixing
45
- - **Info**: Minor suggestions
46
- - **Suggested commit message**: Based on changes
@@ -1,56 +0,0 @@
1
- # Code Review: $ARGUMENTS
2
-
3
- Perform a comprehensive code review following EasyPlatform standards.
4
-
5
- ## Review Scope
6
-
7
- Target: $ARGUMENTS (can be a PR number, file path, or branch name)
8
-
9
- ## Review Checklist
10
-
11
- ### 1. Architecture Compliance
12
- - [ ] Follows Clean Architecture layers (Domain, Application, Persistence, Service)
13
- - [ ] Uses correct repository pattern (I{Service}RootRepository<T>)
14
- - [ ] CQRS pattern: Command/Query + Handler + Result in ONE file
15
- - [ ] No cross-service direct database access
16
-
17
- ### 2. Code Quality
18
- - [ ] Single Responsibility Principle
19
- - [ ] No code duplication
20
- - [ ] Meaningful variable/method names
21
- - [ ] Appropriate error handling with PlatformValidationResult
22
-
23
- ### 3. Platform Patterns
24
- - [ ] Uses platform validation fluent API (.And(), .AndAsync())
25
- - [ ] No direct side effects in command handlers (use entity events)
26
- - [ ] DTO mapping in DTO classes, not handlers
27
- - [ ] Static expressions for entity queries
28
-
29
- ### 4. Security
30
- - [ ] No hardcoded credentials
31
- - [ ] Proper authorization checks
32
- - [ ] Input validation at boundaries
33
- - [ ] No SQL injection risks
34
-
35
- ### 5. Performance
36
- - [ ] Parallel operations where independent
37
- - [ ] Proper pagination for lists
38
- - [ ] No N+1 query patterns
39
- - [ ] Appropriate use of async/await
40
-
41
- ## Output Format
42
-
43
- Provide feedback in this format:
44
-
45
- **Summary:** Brief overall assessment
46
-
47
- **Critical Issues:** (Must fix)
48
- - Issue 1: Description and suggested fix
49
- - Issue 2: Description and suggested fix
50
-
51
- **Suggestions:** (Nice to have)
52
- - Suggestion 1
53
- - Suggestion 2
54
-
55
- **Positive Notes:**
56
- - What was done well
@@ -1,35 +0,0 @@
1
- ---
2
- description: ⚡ Use external agentic tools to scout given directories
3
- argument-hint: [user-prompt] [scale]
4
- ---
5
-
6
- ## Purpose
7
-
8
- Utilize external agentic tools to scout given directories or explore the codebase for files needed to complete the task using a fast, token efficient agent.
9
-
10
- ## Variables
11
-
12
- USER_PROMPT: $1
13
- SCALE: $2 (defaults to 3)
14
- RELEVANT_FILE_OUTPUT_DIR: Use `Report:` from `## Naming` section
15
-
16
- ## Workflow:
17
- - Write a prompt for 'SCALE' number of agents to the `Task` tool that will immediately call the `Bash` tool to run these commands to kick off your agents to conduct the search:
18
- - `gemini -p "[prompt]" --model gemini-2.5-flash-preview-09-2025` (if count <= 3)
19
- - `opencode run "[prompt]" --model opencode/grok-code` (if count > 3 and count < 6)
20
- - if count >= 6, spawn `Explore` subagents to search the codebase in parallel
21
-
22
- **Why use external agentic tools?**
23
- - External agentic tools are faster and more efficient when using LLMs with large context windows (1M+ tokens).
24
-
25
- **How to prompt the agents:**
26
- - If `gemini` or `opencode` is not available, ask the user if they want to install it:
27
- - If **yes**, install it (if there are permission issues, instruct the user to install it manually, including authentication steps)
28
- - If **no**, use the default `Explore` subagents.
29
- - IMPORTANT: Kick these agents off in parallel using the `Task` tool, analyze and divide folders for each agent to scout intelligently and quickly.
30
- - IMPORTANT: These agents are calling OTHER agentic coding tools to search the codebase. DO NOT call any search tools yourself.
31
- - IMPORTANT: That means with the `Task` tool, you'll immediately call the Bash tool to run the respective agentic coding tool (gemini, opencode, claude, etc.)
32
- - IMPORTANT: Instruct the agents to quickly search the codebase for files needed to complete the task. This isn't about a full blown search, just a quick search to find the files needed to complete the task.
33
- - Instruct the subagent to use a timeout of 3 minutes for each agent's bash call. Skip any agents that don't return within the timeout, don't restart them.
34
- - **IMPORTANT:** Sacrifice grammar for the sake of concision when writing reports.
35
- - **IMPORTANT:** In reports, list any unresolved questions at the end, if any.