{"data":{"kind":"file","path":"README.md","version_id":"grn92i0apniiigbeb0mn2mvl","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":1932,"modified_at":"2026-08-10T14:05:52.478000","content_hash":"f9269048f3827574f74931df8c5f05b860ae41d1ce4cd423bdf4f7788fc54331"},"entries":[],"content":"# Secure Coding Challenge\n\nAn evaluation environment that tests an LLM's ability to identify, explain, and fix common web application security vulnerabilities.\n\n## Supported Vulnerability Classes\n\n| Category | Vulnerabilities |\n|----------|----------------|\n| Injection | SQL Injection, Command Injection, XML External Entity (XXE) |\n| Cross-Site Scripting | Reflected XSS, Stored XSS, DOM-based XSS |\n| Access Control | Insecure Direct Object Reference (IDOR), Broken Authentication |\n| Server-Side | Server-Side Request Forgery (SSRF), Path Traversal |\n| Data Handling | Insecure Deserialization, Race Conditions |\n| Client-Side | Open Redirect |\n\n## Evaluation Criteria\n\nEach task is scored across 5 dimensions:\n\n| Dimension | Weight | Description |\n|-----------|--------|-------------|\n| Vulnerability Identification | 30% | Correctly names the vulnerability class |\n| Severity Rating | 10% | Rates severity (LOW/MEDIUM/HIGH/CRITICAL) with justification |\n| Explanation | 20% | Explains why the code is vulnerable and attacker impact |\n| Fixed Code | 25% | Provides a working fix with secure coding patterns |\n| Security Notes | 15% | Lists relevant best practices |\n\n## Task Format\n\nEach task presents a vulnerable code snippet with language and context. The model must respond with structured sections:\n\n1. **Vulnerability Type** - Name the vulnerability class\n2. **Severity** - Rate with justification\n3. **Explanation** - Why it's vulnerable and what an attacker could do\n4. **Fixed Code** - Complete corrected code\n5. **Security Notes** - Best practices\n\n## Languages Covered\n\n- Python (Flask, Django)\n- JavaScript (Express.js, DOM)\n- Java (JDBC, Spring Boot)\n- PHP (mysqli, shell_exec)\n\n## Usage\n\n```python\nimport verifiers as vf\n\nenv = vf.load_environment(\"owner/secure-coding-challenge\")\n# or\nfrom secure_coding_challenge import load_environment\nenv = load_environment()\n```\n\n## Requirements\n\n- `verifiers`\n- `datasets`\n","encoding":"utf-8","truncated":false,"total_bytes":1932},"status":null}