xfi-create-plan
Guide for creating engineering plans through a structured workflow. Use when planning new features, coordinating multi-agent work, or breaking down complex initiatives.
Best use case
xfi-create-plan is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Guide for creating engineering plans through a structured workflow. Use when planning new features, coordinating multi-agent work, or breaking down complex initiatives.
Teams using xfi-create-plan should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/xfi-create-plan/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How xfi-create-plan Compares
| Feature / Agent | xfi-create-plan | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Guide for creating engineering plans through a structured workflow. Use when planning new features, coordinating multi-agent work, or breaking down complex initiatives.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Creating X-Fidelity Engineering Plans
This skill guides the xfi-planner agent through creating comprehensive engineering plans for complex initiatives.
## Prerequisites
- Clear understanding of the feature/initiative to plan
- Access to xfi-system-design agent for architectural input
- Ability to delegate to appropriate subagents for subtask creation
## Workflow Overview
```
┌─────────────────────────────────────────────────────────────┐
│ 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 │
└─────────────────────────────────────────────────────────────┘
```
## Step 1: Requirements Gathering
Ask up to 10 clarifying questions, **one at a time**, building on previous answers.
### Question Categories
1. **Scope**: What is the scope of this initiative?
2. **Goals**: What are the primary goals and success criteria?
3. **Constraints**: Are there any constraints or limitations?
4. **Dependencies**: What existing features/packages does this depend on?
5. **Affected Packages**: Which packages will be impacted?
6. **User Impact**: How will this affect end users?
7. **Testing Strategy**: Any specific testing requirements?
8. **Timeline Sensitivity**: Are there any deadlines or priorities?
9. **Risks**: Any known risks or concerns?
10. **Open Questions**: Anything unclear that needs investigation?
### Question Guidelines
- **STOP after each question** and wait for user response
- **Build on previous answers** - adapt questions based on responses
- **Skip unnecessary questions** if scope is already clear
- **Aim for 5-7 questions** for most initiatives
- **Document key decisions** as you go
## Step 2: Design Consultation
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.
## Step 3: User Confirmation (MANDATORY STOP)
**STOP and present to user:**
```markdown
## 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.**
## Step 4: Create Plan Directory and Index File
Once confirmed, create the plan structure:
### Directory Structure
```bash
knowledge/plans/[YYYYMMDD]-[feature-name]/
```
### Index File Template
Create `index-[feature-name]-[yyyymmdd].md`:
```markdown
# 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
```
## Execution Order
### Phase 1 (Parallel)
| Subtask | Subagent | Description | Dependencies |
|---------|----------|-------------|--------------|
| 01 | xfi-engineer | [Description] | None |
| 02 | xfi-engineer | [Description] | None |
### Phase 2 (After Phase 1)
| Subtask | Subagent | Description | Dependencies |
|---------|----------|-------------|--------------|
| 03 | xfi-engineer | [Description] | 01, 02 |
### Phase 3 (After Phase 2)
| Subtask | Subagent | Description | Dependencies |
|---------|----------|-------------|--------------|
| 04 | xfi-testing-expert | [Description] | 03 |
## Global Definition of Done
- [ ] All subtasks completed
- [ ] All unit tests passing
- [ ] Code review by xfi-code-reviewer completed
- [ ] Architecture review by xfi-system-design completed
- [ ] Global test suite passing (verified by xfi-testing-expert)
- [ ] Documentation updated by xfi-docs-expert
- [ ] Knowledge captured by xfi-system-design
- [ ] User approval received
## Execution Notes
[To be filled during plan execution]
## Completion Checklist
- [ ] Subtask 01: Pending
- [ ] Subtask 02: Pending
- [ ] Subtask 03: Pending
- [ ] Subtask 04: Pending
- [ ] Final code review
- [ ] Final architecture review
- [ ] Global tests verified
- [ ] Documentation updated
- [ ] Knowledge captured
```
## 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**:
- Only create/edit the subtask markdown file
- Do NOT edit any code files
- Do NOT run any tests
- Include instruction in the file that executing agent should NOT trigger global tests
**Subtask Details**:
- Feature: [Feature name]
- Subtask: [Subtask name]
- Objective: [Clear objective]
- Deliverables: [Expected deliverables]
- Dependencies: [List of dependent subtask IDs or None]
**File Template**:
[Include the subtask file template from xfi-planner agent documentation]
After creating the file, report back with:
1. File path created
2. Summary of deliverables defined
3. Any concerns or questions about the subtask"
```
### 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:
1. Clear and achievable deliverables
2. Appropriate subagent assignments
3. Correct dependency ordering
4. Complete definition of done criteria
5. Any gaps or concerns
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:
1. Alignment with X-Fidelity architecture
2. Appropriate package assignments
3. Completeness of scope
4. Technical feasibility
5. Risk coverage
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?
```
## Step 8: Finalize Plan
Once user approves:
1. Update index file status to "Ready for Review"
2. Document execution instructions:
```markdown
## 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
```
## Critical Rules
1. **NEVER edit code files** during planning
2. **ALWAYS stop for user confirmation** at Steps 3 and 7
3. **DELEGATE subtask creation** to appropriate subagents
4. **INCLUDE testing restrictions** in all subtask files
5. **CAPTURE all decisions** in the index file
6. **UPDATE index** with discoveries from subagents
## Troubleshooting
### Subagent doesn't create correct file
- Verify the file path is correct
- Ensure template was provided
- Re-invoke with clearer instructions
### Scope too large
- Break into multiple plans
- Phase the implementation
- Prioritize critical subtasks
### Unclear requirements
- Ask additional clarifying questions
- Consult xfi-system-design for guidance
- Document assumptions clearlyRelated Skills
u0048-planning-dependency-dag-planner
Build and operate the "Planning Dependency DAG Planner" capability for Strategic Planning and Decomposition. Use when outcomes in this capability family are required for production execution.
skill-create-flow
Create new high-quality agent skills with a standalone, repeatable workflow (no dependency on other skills). Use when you want to go from a vague skill idea → narrow scope → extract expert frameworks → write SKILL.md + examples/evals/index/changelog artifacts → self-validate with test prompts.
remember-plan
Remember a plan with git commit
planning
Provides expertise on how to plan for work in this repo
plan-generator
Creates structured plans from requirements. Generates comprehensive plans with steps, dependencies, risks, and success criteria. Coordinates with specialist agents for planning input and validates plan completeness. Uses template-renderer for formatted output.
plan-first-development
Break complex projects into detailed multi-phase plans before writing any code. Prevents scope creep, refactoring waste, and missed requirements. Use when starting any project with more than 1000 lines of code or building systems with multiple interacting components.
Plan Dependency Graph
Build and analyze dependency graphs across plans — identify blocking dependencies, critical paths, parallel work opportunities, and stale plans
idea-plan
科研idea规划与跨对话追踪skill。将科研需求转化为 Milestone-Keypoint 两级任务体系,以 .plan 文件结构支持跨对话状态追踪。触发场景:用户需要规划实施路径、追踪多阶段进度、跨对话继续科研项目、调整现有计划。
git-create-commit-skill
Use when creating git commits - mandates commits format with required scopes for all new commits, provides scope discovery from codebase structure, and prevents rationalization under pressure (emergency, fatigue, ambiguity)
family-history-planning
Provides assistance with planning family history and genealogy research projects.
create-target
Scaffold a new target type (e.g., Zed, Windsurf) for universal-ai-config
create-project
Initialize new projects with CAMI capabilities and recommended agent teams. Triggers on "create a new project", "start a new project", "initialize project", "set up project with agents", "what agents do I need for this project", "help me set up my project roster". Guides through project requirements, tech stack detection, agent team recommendations, and full CAMI initialization.