What problem does it solve? Black-box testing alone misses vulnerabilities hidden in internal routes, cron jobs, and message handlers. This Skill performs white-box static analysis with Semgrep MCP to locate vulnerable code patterns, then guides dynamic validation so findings become reportable, evidence-backed vulnerabilities. ## Core Features & Use Cases - Broad and Class-Specific Scanning: Run configs like p/security-audit, p/sql-injection, p/xss, p/ssrf, p/secrets, and p/jwt to sweep codebases for OWASP Top 10 patterns across supported languages. - Custom Rule Authoring: Write YAML rules for app-specific patterns such as missing auth decorators, f-string SQL injection, and open redirects. - Taint and AST Analysis: Use get_abstract_syntax_tree to trace user input to dangerous sinks across files, decorators, and middleware chains. - Use Case: After decompiling an APK with jadx, scan the Java source with Semgrep to find hardcoded API keys and exported components, then validate each finding with a real HTTP request before reporting. ## Quick Start Scan this repository with Semgrep for SQL injection and hardcoded secrets, then list the reachable sinks I should validate dynamically.