What problem does it solve? Web applications built on Spring Boot or legacy JSP/Servlet stacks often contain authentication and session weaknesses—hardcoded JWT secrets, plaintext password storage, session fixation, and missing cookie security flags—that static scanners alone either miss or over-report. This Skill combines Semgrep rule scanning with AI context verification to confirm real vulnerabilities and filter out false positives. ## Core Features & Use Cases - Stack-aware scanning: Automatically detects spring-modern vs jsp-legacy codebases and applies the matching Semgrep ruleset, with a grep fallback when Semgrep is unavailable. - AI context verification: Each scanner candidate is verified against actual code—JWT signature validation, session regeneration on login, BCrypt usage, and cookie Secure/HttpOnly/SameSite attributes. - Structured severity report: Confirmed findings are reported with what is vulnerable, why, how it can be exploited, and how to fix it, mapped to CWE and OWASP A07:2021. - Use Case: Point it at a legacy eGovFrame project to confirm that hash="plaintext" in context-security.xml means passwords are stored as reversible Base64, and receive a BCrypt migration fix. ## Quick Start Ask the AI to check the target project source for authentication, session, and JWT vulnerabilities using this skill.