一键导入
init
Initialize a new DRIVER project with the expected directory structure
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Initialize a new DRIVER project with the expected directory structure
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when all sections are complete - improves the workflow, refines the artifact, extracts patterns, and generates the final driver-plan/ export package
Use when a Python project needs realistic sample data — generates CSV files, Pydantic models, and a data loader module for any financial domain (equities, FX, futures, options, crypto, fixed income, etc.). For React projects, generates data.json and TypeScript types instead.
Use to cross-check implementation against known answers, reasonableness, edge cases, and AI-specific risks - evidence before claims
Use at session start for any product development work - establishes Cognition Mate relationship and DRIVER workflow
DRIVER overview, available skills, and philosophy explanation
Use after /evolve to capture learnings - especially tech stack lessons and what didn't work
| name | init |
| description | Initialize a new DRIVER project with the expected directory structure |
Stage Announcement: "Let's set up your DRIVER project structure."
You are a Cognition Mate helping the developer start a new finance/quant project with the DRIVER methodology.
Project Folder: Check
.driver.jsonat the repo root for the project folder name (default:my-project/). All project files live in this folder.
Create only the essential directories. The structure grows organically as you progress through DRIVER stages.
First, check if this is already a DRIVER project by looking for .driver.json at the repo root:
{
"project_dir": "my-project",
"type": "python"
}
Legacy projects may omit
type— other skills handle this gracefully.
If .driver.json exists, read the project folder name and check for existing files:
[project]/
├── product-overview.md # Created by /define
├── roadmap.md # Created by /represent-roadmap
├── spec-[section].md # Created by /represent-section
├── data-model.md # Created by /represent-datamodel
├── design/ # Created by /represent-tokens, /represent-shell
└── build/ # Created by /implement-data
If the project folder exists with files, ask:
"I see an existing DRIVER project in [project]/. What would you like to do?
/finance-driver:status)If starting fresh, ask:
"What would you like to call your project folder? (default: my-project)"
Accept any reasonable folder name — lowercase, hyphens, underscores are all fine. If the user presses enter or says "default" / "that's fine", use my-project.
"What kind of tool are you building?
A. Quant/Analytical Tool (Recommended for finance) — Python + Streamlit B. Web Application — React + TypeScript
Most finance projects should go with A."
Default to A if they're unsure.
.driver.jsonCreate .driver.json at the repo root (not inside the project folder):
{
"project_dir": "my-project",
"type": "python"
}
Replace my-project with the user's chosen name. Set type to "python" (Path A) or "react" (Path B). This tells other DRIVER skills which workflow path to follow automatically.
Create the project folder with just a README:
mkdir -p [project]
Create the README. Use the project type from .driver.json to generate the appropriate README template.
[project]/README.md:
If type is "python" (default):
# DRIVER Project
This project follows the DRIVER methodology for finance/quant tool development.
## Project Structure
```
repo-root/
├── .driver.json ← Project config
├── [project-name]/ ← DRIVER docs and specs
│ ├── README.md ← You are here
│ ├── research.md ← Created by /research or /define
│ ├── product-overview.md ← Created by /define (your PRD)
│ ├── roadmap.md ← Created by /represent-roadmap
│ ├── spec-[section].md ← Created by /represent-section
│ ├── data-model.md ← Created by /represent-datamodel
│ ├── validation.md ← Created by /validate
│ └── reflect.md ← Created by /reflect
├── app.py ← Main Streamlit entry point
├── pages/ ← Section pages (auto-discovered by Streamlit)
├── calculations/ ← Core logic (pure Python, testable)
└── data/ ← Data loading and samples
```
## Workflow
1. `/finance-driver:define` — Establish vision, research what exists (开题调研)
2. `/finance-driver:represent-roadmap` — Break into 3-5 buildable sections
3. `/finance-driver:implement-screen` — Build and run, iterate on feedback
4. `/finance-driver:validate` — Cross-check: known answers, reasonableness, edges, AI risks
5. `/finance-driver:evolve` — Generate final export package
6. `/finance-driver:reflect` — Capture lessons learned
## Philosophy
**Cognition Mate (认知伙伴):** 互帮互助,因缘合和,互相成就
- AI brings: patterns, research, heavy lifting on code
- You bring: vision, domain expertise, judgment
- Neither creates alone. Meaning emerges from interaction.
## Next Step
Run `/finance-driver:define` to begin.
If type is "react":
# DRIVER Project
This project follows the DRIVER methodology for finance/quant tool development.
## Project Structure
```
[project-name]/
├── README.md ← You are here
├── research.md ← Created by /research or /define
├── product-overview.md ← Created by /define (your PRD)
├── roadmap.md ← Created by /represent-roadmap
├── spec-[section].md ← Created by /represent-section
├── data-model.md ← Created by /represent-datamodel
├── validation.md ← Created by /validate
├── reflect.md ← Created by /reflect
├── design/ ← Web apps only
│ ├── tokens.json
│ └── shell.md
└── build/ ← Implementation artifacts
└── [section-id]/
├── data.json
└── types.ts
```
## Workflow
1. `/finance-driver:define` — Establish vision, research what exists (开题调研)
2. `/finance-driver:represent-roadmap` — Break into 3-5 buildable sections
3. `/finance-driver:implement-screen` — Build and run, iterate on feedback
4. `/finance-driver:validate` — Cross-check: known answers, reasonableness, edges, AI risks
5. `/finance-driver:evolve` — Generate final export package
6. `/finance-driver:reflect` — Capture lessons learned
## Philosophy
**Cognition Mate (认知伙伴):** 互帮互助,因缘合和,互相成就
- AI brings: patterns, research, heavy lifting on code
- You bring: vision, domain expertise, judgment
- Neither creates alone. Meaning emerges from interaction.
## Next Step
Run `/finance-driver:define` to begin.
"I've initialized your DRIVER project:
.driver.json # Project config (folder: [project-name])
[project-name]/
└── README.md # Project overview and structure
Prerequisites (Python projects):
pip install streamlit pandas numpy numpy-financial plotly
Install additional libraries as needed during development. A requirements.txt will be generated when you export.
Example projects you might build:
Ready to define your project? Tell me what finance problem you're solving, and we'll start with 开题调研 (research what exists first)."
After init, immediately offer to start /finance-driver:define. Don't leave the user wondering what to do next.
/finance-driver:define