OpenTaint

OpenTaint is an open source taint analysis engine for application security. It is designed to find issues that AST-pattern matchers miss and to let LLM agents turn discovered vulnerabilities into reusable rules. The page describes it as a formal substrate for AI application security, aimed at problems created by AI-generated production code and by the limits of non-deterministic security scanning.

Its analysis engine tracks untrusted data across function boundaries, persistence layers, aliases, and async code. OpenTaint uses AST-pattern rules, and it applies those rules across the entire codebase deterministically in minutes of CPU time. The page says it includes more than 100 rules across more than 20 vulnerability classes, with examples such as SQL injection, XSS, SSRF, SpEL injection, open redirects, path traversal, and command injection. Reports trace findings from the HTTP source through method calls, async boundaries, and JPA persistence to the dangerous call, and tie the result back to the Spring endpoint where the data entered.

OpenTaint supports Java and Kotlin today. For those languages, the analysis runs on bytecode, which the page says gives precise resolution of inheritance, generics, and library calls, including the standard library and third-party JARs in the build classpath. Spring Boot is supported deeply, including Spring MVC and Spring Data, and the page also names Reactor, Spring WebFlux, and Kotlin coroutines in its flow modeling. The site presents OpenTaint through the web, with a CLI and an Agent Viewer, and it provides installation commands for npm, brew, PowerShell, Docker, and curl-based setup.

The stack ships under Apache 2.0 and MIT. The page says there is no paid tier to unlock taint tracking and no gate on writing custom rules.