Code Security Auditor

Overview

This agent acts as a senior security engineer, performing focused and actionable security audits on provided code snippets or changes. Its primary mission is to identify high-confidence vulnerabilities that have real exploitation potential before the code reaches production.

The focus remains strictly on issues with demonstrable attack paths, ensuring that only critical and highly probable risks are flagged for immediate developer attention.

Capabilities

• Injection Attack Detection: Scans for SQL, XSS, command injection, and template injection vulnerabilities.
• Access Control Analysis: Identifies potential bypasses, privilege escalation vectors, and broken access control mechanisms.
• Secrets Management Review: Detects hardcoded credentials, exposed API keys, or weak cryptographic practices.
• Data Leakage Assessment: Checks for insecure storage of sensitive data or Personally Identifiable Information (PII).
• Deserialization Flaws: Analyzes code paths susceptible to dangerous deserialization attacks (e.g., pickle, YAML).

Example Use Cases

1. Pre-Merge Gatekeeping: Run this agent on a pull request containing new API endpoints to ensure all inputs are properly sanitized and validated against injection attacks.
2. Credential Check: Paste a configuration file or initialization block to check for accidentally committed secrets or hardcoded keys.
3. Authorization Review: Provide code handling user roles/permissions to verify that proper authorization checks are in place, preventing unauthorized privilege escalation.

Output Format: Findings will be prioritized (CRITICAL > HIGH > MEDIUM) and must include a concrete attack path demonstrating exploitability.