Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/mrchevyceleb/edgetest-skill

Comprehensive edge case testing for Claude Code with auto-fix and production deployment
https://github.com/mrchevyceleb/edgetest-skill

ai automation claude-code cli developer-tools edge-cases playwright testing

Last synced: 19 days ago
JSON representation

Comprehensive edge case testing for Claude Code with auto-fix and production deployment

Awesome Lists containing this project

README 

          
# /edgetest - Comprehensive Edge Case Testing for Claude Code



**Automated edge case testing with auto-fix loop and production deployment verification**



[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)

[![Claude Code](https://img.shields.io/badge/Claude%20Code-Skill-blue)](https://github.com/anthropics/claude-code)



---



## Overview



`/edgetest` is a Claude Code skill that automatically discovers, tests, and fixes edge cases in web applications. Unlike traditional testing tools, it:



- 🔍 **Auto-discovers** edge cases across 8 categories

- 🔧 **Auto-fixes** failures with intelligent root cause analysis

- 🚀 **Deploys to production** and retests in live environment

- 🤖 **Spawns background agents** for complex issues

- ✅ **Requires 100% pass** (binary, not scored)

- 🔐 **Handles authentication** automatically with session restoration



---



## Features



### 8 Edge Case Categories



1. **Input Validation** - SQL injection, XSS, Unicode, max length, special chars

2. **Data States** - Empty, single item, pagination, large datasets, missing fields

3. **Network/API Errors** - 404, 500, timeouts, rate limits, CORS, offline

4. **UI/Interaction** - Rapid clicks, double submit, back button, resize, focus/blur

5. **Auth/Authorization** - Session expiry, invalid tokens, permission changes

6. **Browser/Device** - Mobile/tablet/desktop, zoom, dark mode, navigation

7. **Timing/Race Conditions** - Delayed scripts, concurrent calls, stale data

8. **Boundary Values** - Min/max, off-by-one, zero, negative, decimals



### Auto-Fix Loop



When a test fails:

1. Analyzes root cause

2. Implements fix

3. Validates build

4. **Deploys to production**

5. **Retests in live environment**

6. Repeats up to 3 times

7. Creates background task if unresolved



### Authentication Handling



- **First run:** Prompts user to login, saves session automatically

- **Subsequent runs:** Restores session (localStorage, sessionStorage, cookies)

- **Session expiry:** Detects 401/403, prompts for re-auth, saves new session

- **Secure:** Session files gitignored, stored locally only



---



## Installation



### Prerequisites



1. **[Claude Code CLI](https://github.com/anthropics/claude-code)** installed

2. **[Playwright MCP](https://github.com/modelcontextprotocol/servers/tree/main/src/playwright)** configured

3. **Git** repository initialized

4. **Deployment platform** (Vercel, Railway, or custom)



### Quick Install



**macOS/Linux:**

```bash

curl -fsSL https://raw.githubusercontent.com/mrchevyceleb/edgetest-skill/main/install.sh | bash

```



**Windows (PowerShell):**

```powershell

iwr -useb https://raw.githubusercontent.com/mrchevyceleb/edgetest-skill/main/install.ps1 | iex

```



### Manual Install



1. **Download the skill file:**

   ```bash

   curl -o ~/.claude/commands/edgetest.md https://raw.githubusercontent.com/mrchevyceleb/edgetest-skill/main/edgetest.md

   ```



2. **Restart Claude Code** to load the skill



3. **Verify installation:**

   ```bash

   # The skill should appear in autocomplete

   /edgetest --help

   ```



---



## Usage



### Basic Syntax



```bash

/edgetest [focus] [url] [options]

```



### Arguments



| Argument | Description | Example |

|----------|-------------|---------|

| `focus` | Scope testing to specific feature (optional) | `email integration`, `payment flow` |

| `url` | Production URL to test (optional) | `https://myapp.vercel.app` |



### Options



| Option | Description | Default |

|--------|-------------|---------|

| `--categories=cat1,cat2` | Test only specific categories | All 8 categories |

| `--severity=level` | Test only specific severity | All severities |

| `--max-fixes=N` | Max auto-fix attempts per test | 3 |

| `--skip-deploy` | Skip production deployment | false |

| `--config=path` | Custom edge case config file | None |



---



## Examples



### Example 1: Full App Testing



```bash

/edgetest https://myapp.vercel.app

```



**Result:** Tests entire app (~50-100 tests), auto-fixes failures, deploys to production



---



### Example 2: Focused Testing



```bash

/edgetest email integration https://myapp.vercel.app

```



**Result:** Tests only email-related features (~10-30 tests), faster iteration



---



### Example 3: Critical Tests Only



```bash

/edgetest --severity=critical

```



**Result:** Tests only CRITICAL edge cases (SQL injection, XSS, auth bypass, etc.)



---



### Example 4: Local Testing (No Deploy)



```bash

/edgetest --skip-deploy

```



**Result:** Tests edge cases locally without deploying to production



---



## How It Works



### Workflow



```

1. SCOPE DETECTION

   ↓ Determine full app vs focused testing



2. DISCOVERY

   ↓ Map forms, inputs, APIs, pages

   ↓ Generate ~10-100 edge case tests

   ↓ Prioritize by severity



3. SYSTEMATIC TESTING

   ↓ Execute tests with Playwright

   ↓ IF FAIL → Auto-Fix Loop



4. AUTO-FIX LOOP

   ↓ Analyze root cause

   ↓ Implement fix

   ↓ Validate build



5. PRODUCTION DEPLOYMENT

   ↓ Git commit + push

   ↓ Deploy (Vercel/Railway/custom)

   ↓ Wait 2 minutes



6. RE-TEST IN PRODUCTION

   ↓ Re-run failed test

   ↓ IF PASS → Next test

   ↓ IF FAIL (< 3 attempts) → Try different fix

   ↓ IF FAIL (>= 3 attempts) → Spawn background agent



7. COMPLETION

   ↓ Generate comprehensive report

```



---



## Comparison with Other Testing Tools



| Feature | /edgetest | /bigtest | /perfect | Jest | Playwright |

|---------|-----------|----------|----------|------|------------|

| **Focus** | Edge cases only | UI/UX audit | Feature impl | Unit tests | E2E tests |

| **Auto-discovery** | ✅ | ❌ | ❌ | ❌ | ❌ |

| **Auto-fix** | ✅ | ❌ | ✅ | ❌ | ❌ |

| **Production deploy** | ✅ | ❌ | ❌ | ❌ | ❌ |

| **Auth handling** | ✅ Auto | Manual | Manual | Manual | Manual |

| **Pass criteria** | 100% binary | 0-100 score | 0-100 score | % coverage | Manual |



---



## Authentication



For apps requiring login, `/edgetest` handles authentication automatically:



### First Run

1. Navigate to app

2. Detect login requirement

3. **Prompt user to login**

4. Save session to `.edgetest/auth-session.json`

5. Proceed with testing



### Subsequent Runs

1. Navigate to app

2. **Auto-restore session** (no login needed!)

3. Proceed with testing



### Session Expiry

1. Detect 401/403 errors

2. Delete expired session

3. Prompt for re-auth

4. Save new session



**Security:** Session files are gitignored and stored locally only.



---



## Output



### Final Report



```markdown

===============================================================================

                      /EDGETEST EXECUTION REPORT

===============================================================================



APP: MyApp

PRODUCTION URL: https://myapp.vercel.app

STATUS: ✅ 100% PASSING



SUMMARY:

  Total Tests: 87

  Passed: 87 (100%)

  Failed: 0 (0%)



Auto-Fixes:

  ✅ Resolved: 12

  ⚠️  Background Tasks: 0



Deployments: 12

Total Execution Time: 1h 23m 15s



===============================================================================

```



---



## Configuration



### Custom Edge Cases



Create `edgetest-config.json`:



```json

{

  "categories": {

    "input": {

      "enabled": true,

      "customTests": [

        {

          "name": "Credit card validation",

          "input": "4111 1111 1111 1111",

          "expected": "Accepted or formatted"

        }

      ]

    }

  },

  "exclusions": [

    "/admin/*",

    "/debug/*"

  ]

}

```



Use:

```bash

/edgetest --config=edgetest-config.json

```



---



## Troubleshooting



### Issue: Playwright MCP not available



**Solution:**

```bash

# Check .mcp.json or ~/.claude.json for playwright config

# Restart Claude Code to reload MCPs

```



### Issue: Deployment not detected



**Solution:**

Add `vercel.json`, `railway.json`, or deploy script to `package.json`:

```json

{

  "scripts": {

    "deploy": "your-deploy-command"

  }

}

```



### Issue: Session expired



**Solution:**

```bash

# Delete session file to force fresh login

rm .edgetest/auth-session.json

```



---



## Best Practices



### 1. Run after implementing features



```bash

# Implement feature first

/perfect implement payment flow



# THEN test edge cases

/edgetest payment flow

```



### 2. Use focus parameter for faster iteration



```bash

# Focused testing (10-30 tests, faster)

/edgetest email integration



# Full app testing (50-100 tests, comprehensive)

/edgetest

```



### 3. Review background tasks promptly



After `/edgetest` completes, check `/tasks/` for background tasks and fix manually.



### 4. Combine with other skills



**Recommended workflow:**

1. `/perfect` - Implement feature (100/100 score)

2. `/edgetest` - Test edge cases (deploy fixes)

3. `/bigtest` - Full app audit (periodically)



---



## FAQ



### Q: How is /edgetest different from /bigtest?



**A:** `/bigtest` is a comprehensive UI/UX audit (0-100 score). `/edgetest` focuses ONLY on edge cases, auto-fixes failures, and deploys to production (100% binary pass).



### Q: Will it break my production app?



**A:** No. Each fix is build-validated, tested locally, and deployed incrementally. If re-test fails, the loop tries a different approach.



### Q: What if auto-fix doesn't work?



**A:** After 3 failed attempts, a background task is created with full context, attempted fixes, and recommended next steps.



### Q: Does it work with all frameworks?



**A:** Yes! Framework-agnostic (React, Vue, Angular, Svelte, etc.). Works with any web app accessible via URL.



---



## Contributing



Contributions welcome! Please read [CONTRIBUTING.md](CONTRIBUTING.md) first.



### Development Setup



1. Clone the repository

2. Edit `edgetest.md` (the skill file)

3. Test locally by copying to `~/.claude/commands/edgetest.md`

4. Submit a pull request



---



## License



MIT License - see [LICENSE](LICENSE) file for details.



---



## Support



- **Issues:** [GitHub Issues](https://github.com/mrchevyceleb/edgetest-skill/issues)

- **Discussions:** [GitHub Discussions](https://github.com/mrchevyceleb/edgetest-skill/discussions)

- **Documentation:** Full guide in `edgetest.md`



---



## Acknowledgments



- Built for [Claude Code CLI](https://github.com/anthropics/claude-code)

- Uses [Playwright MCP](https://github.com/modelcontextprotocol/servers/tree/main/src/playwright) for browser automation

- Inspired by the need for comprehensive edge case testing in production apps



---



**Made with ❤️ by [Matt Johnston](https://github.com/mrchevyceleb)**



**Star ⭐ this repo if you find it useful!**