Skip to content

AGENTS.md

Latest commit

 

History

History
101 lines (78 loc) · 2.73 KB

AGENTS.md

File metadata and controls

101 lines (78 loc) · 2.73 KB
Raw

AGENTS.md

This file provides guidance to AI coding agents (Claude Code, Cursor, Copilot, etc.) when working with code in this repository.

Repository Overview

A collection of skills for AI coding agents covering Stripe integration and Tauri desktop development. Skills are packaged instructions, rules, and reference documentation that extend agent capabilities with deep domain expertise.

Directory Structure

drivers/
  stripe-expert/            # Stripe product advisor (start here for Stripe)
  stripe-dev/               # Stripe integration orchestrator
  stripe-payments/          # Payment acceptance
  stripe-billing/           # Subscriptions & invoicing
  stripe-connect/           # Marketplace & platforms
  stripe-treasury/          # Embedded finance
  stripe-revenue/           # Tax, recognition, reporting
  stripe-money-management/  # Payouts, refunds, disputes, fraud
  tauri-dev/                # Tauri v2 desktop development
  dist/                     # Packaged .skill files

Per-Skill Structure

{skill-name}/
  SKILL.md              # Required: skill definition with frontmatter
  metadata.json         # Skill metadata (version, references)
  rules/                # Individual rule files (concise best practices)
    _sections.md        # Section ordering and impact levels
    _template.md        # Template for new rules
    {prefix}-{name}.md  # Rule files with YAML frontmatter
  references/           # Deep-dive reference documentation
    {topic}.md          # Comprehensive guides with code examples

SKILL.md Format

---
name: {skill-name}
description: "What the skill does and when to trigger it"
license: MIT
metadata:
  author: drivers
  version: "1.0.0"
---

Rule File Format

---
title: Rule Title
impact: CRITICAL | HIGH | MEDIUM | LOW
impactDescription: "Brief impact description"
tags: tag1, tag2, tag3
---

## Rule Title

Brief explanation.

**Incorrect (what's wrong):**
```typescript
// Bad code

Correct (what's right):

// Good code

Reference: Link


### Impact Levels

- `CRITICAL` - Must follow, major bugs or security issues if ignored
- `HIGH` - Strongly recommended, significant quality improvement
- `MEDIUM` - Best practice, moderate improvement
- `LOW` - Nice to have, incremental improvement

## End-User Installation

```bash
# Claude Code
cp -r {skill-name} ~/.claude/skills/

# claude.ai
# Add SKILL.md contents to project knowledge

Context Efficiency

  • SKILL.md is loaded when skill triggers (~100-200 lines)
  • Rules are loaded individually as needed (~30-50 lines each)
  • References are loaded only for deep dives (~500-1500 lines each)
  • Progressive disclosure: metadata -> SKILL.md -> rules -> references