القائمة
Guide for creating engineering plans through a structured workflow. Use when planning new features, coordinating multi-agent work, or breaking down complex initiatives.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Multi-purpose utility for CRUX compression workflows. Provides token estimation and checksum calculation. Use when estimating tokens for compression, comparing file sizes, or getting checksums for sourceChecksum tracking.
Guide for executing engineering plans through coordinated subagent work. Use when executing existing plans from knowledge/plans/ directory.
Guide for creating a new package in the X-Fidelity monorepo. Use when adding new packages, setting up monorepo structure, or configuring workspace dependencies.
Guide for ensuring CLI and VSCode extension produce identical analysis results. Use when verifying CLI-Extension parity, debugging output differences, or setting up consistency checks.
Guide for creating a new X-Fidelity archetype configuration. Use when defining project templates, configuring rule sets, or setting up dependency requirements.
Guide for creating a new X-Fidelity plugin from the template. Use when creating plugins, adding new analysis capabilities, or extending the plugin system.
| name | xfi-create-plan |
| description | Guide for creating engineering plans through a structured workflow. Use when planning new features, coordinating multi-agent work, or breaking down complex initiatives. |
This skill guides the xfi-planner agent through creating comprehensive engineering plans for complex initiatives.
┌─────────────────────────────────────────────────────────────┐
│ 1. Requirements Gathering (up to 10 questions) │
├─────────────────────────────────────────────────────────────┤
│ 2. Design Consultation (xfi-system-design agent) │
├─────────────────────────────────────────────────────────────┤
│ 3. USER STOP: Confirm decisions │
├─────────────────────────────────────────────────────────────┤
│ 4. Create plan directory and index file │
├─────────────────────────────────────────────────────────────┤
│ 5. Delegate subtask file creation to subagents │
├─────────────────────────────────────────────────────────────┤
│ 6. Review by xfi-code-reviewer and xfi-system-design │
├─────────────────────────────────────────────────────────────┤
│ 7. USER STOP: Final review │
├─────────────────────────────────────────────────────────────┤
│ 8. Declare plan ready with execution instructions │
└─────────────────────────────────────────────────────────────┘
Ask up to 10 clarifying questions, one at a time, building on previous answers.
After gathering requirements, consult the xfi-system-design agent:
Invoke xfi-system-design subagent with:
"Provide architectural input for this initiative:
[Summary of requirements gathered]
Key questions:
1. What packages need to be modified?
2. What is the recommended architecture approach?
3. Are there any design patterns we should follow?
4. What are the main technical risks?
5. Suggest a logical breakdown of subtasks"
Capture the design agent's recommendations.
STOP and present to user:
## Plan Summary for Confirmation
### Initiative: [Name]
### Key Decisions Made:
1. [Decision 1 with rationale]
2. [Decision 2 with rationale]
3. [Decision 3 with rationale]
### Packages Affected:
- [package 1] - [changes needed]
- [package 2] - [changes needed]
### Proposed Subtasks:
1. [Subtask 1] - assigned to [subagent]
2. [Subtask 2] - assigned to [subagent]
### Architecture Approach:
[Summary from xfi-system-design]
### Risks Identified:
- [Risk 1]
- [Risk 2]
---
Do you approve these decisions? Any changes or questions?
Wait for user confirmation before proceeding.
Once confirmed, create the plan structure:
knowledge/plans/[YYYYMMDD]-[feature-name]/
Create index-[feature-name]-[yyyymmdd].md:
# Plan: [Feature Name]
## Status
Draft
## Overview
[High-level description from requirements]
## Key Decisions
- [Decision 1]: [Rationale]
- [Decision 2]: [Rationale]
## Requirements
[Numbered list of requirements gathered]
## Design Notes
[Summary from xfi-system-design consultation]
## Subtask Dependency Graph
```mermaid
graph TD
subgraph "Phase 1 (Parallel)"
A[subtask-01-setup]
B[subtask-02-types]
end
subgraph "Phase 2"
C[subtask-03-implementation]
end
subgraph "Phase 3"
D[subtask-04-testing]
end
A --> C
B --> C
C --> D
| Subtask | Subagent | Description | Dependencies |
|---|---|---|---|
| 01 | xfi-engineer | [Description] | None |
| 02 | xfi-engineer | [Description] | None |
| Subtask | Subagent | Description | Dependencies |
|---|---|---|---|
| 03 | xfi-engineer | [Description] | 01, 02 |
| Subtask | Subagent | Description | Dependencies |
|---|---|---|---|
| 04 | xfi-testing-expert | [Description] | 03 |
[To be filled during plan execution]
## Step 5: Delegate Subtask File Creation
For each subtask, invoke the appropriate subagent to create the subtask file.
### Delegation Template
Invoke [subagent-name] subagent with:
"Create a subtask plan file for the following task.
File to create: knowledge/plans/[yyyymmdd]-[feature-name]/subtask-[NN]-[feature]-[subtask-name]-[yyyymmdd].md
IMPORTANT:
Subtask Details:
File Template: [Include the subtask file template from xfi-planner agent documentation]
After creating the file, report back with:
### Subagent Selection Guide
| Task Type | Subagent |
|-----------|----------|
| General implementation | xfi-engineer |
| Plugin development | xfi-plugin-expert |
| VSCode extension | xfi-vscode-expert |
| Rule/archetype work | xfi-rules-expert |
| Testing strategy | xfi-testing-expert |
| Security-sensitive | xfi-security-expert |
| Build/CI work | xfi-build-expert |
### Collect Results
As each subagent completes:
1. Verify the subtask file was created correctly
2. Update the index file if any new information was discovered
3. Note any concerns raised by subagents
## Step 6: Plan Review
Invoke reviewers in parallel:
### xfi-code-reviewer
Invoke xfi-code-reviewer subagent with:
"Review the engineering plan files for completeness and quality:
Plan directory: knowledge/plans/[yyyymmdd]-[feature-name]/
Check for:
Provide suggestions for improvements."
### xfi-system-design
Invoke xfi-system-design subagent with:
"Review the engineering plan for architectural soundness:
Plan directory: knowledge/plans/[yyyymmdd]-[feature-name]/
Check for:
Provide suggestions for improvements."
### Incorporate Feedback
Update plan files based on reviewer suggestions.
## Step 7: User Final Review (MANDATORY STOP)
**STOP and present to user:**
```markdown
## Plan Ready for Review
### Plan Location
`knowledge/plans/[yyyymmdd]-[feature-name]/`
### Files Created
- `index-[feature-name]-[yyyymmdd].md` (main plan)
- `subtask-01-[feature]-[subtask-name]-[yyyymmdd].md`
- `subtask-02-[feature]-[subtask-name]-[yyyymmdd].md`
- ...
### Review Feedback Incorporated
- [Summary of changes from code reviewer]
- [Summary of changes from system design]
### Execution Summary
- Total subtasks: [N]
- Parallel phases: [N]
- Estimated complexity: [Low/Medium/High]
---
Please review the plan files. When ready to execute, use:
`/xfi-plan-execute`
Any changes needed before finalizing?
Once user approves:
## Execution Instructions
To execute this plan in a new agent session:
1. Start a new conversation
2. Run: `/xfi-plan-execute`
3. Select this plan: `[yyyymmdd]-[feature-name]`
4. Confirm execution summary
5. Monitor subtask progress
6. Review final changes when prompted
7. Approve documentation updates