Sloppoke

The tool identifies residues such as defensive caps, hardcoded secrets, AI-style comments, and dead code stubs that often slip past human review. It processes staged changes locally before they land in the repository and can either strip problematic sections or insert TODO markers for later attention. Adaptive learning allows the detector to refine its rules based on user feedback over time, reducing false positives while catching new failure modes. It supports multiple platforms and integrates with popular editors and agents through simple command hooks. Unlike traditional linters or post-push review services, Sloppoke emphasizes speed and privacy with a focus on statistical correlation between LLM patterns and known codebase problems rather than runtime guarantees.