Dead Code Detector is a skill that performs static analysis across multi-language codebases to identify code that is no longer in use. It cross-references symbols across files rather than analyzing a single file in isolation, so it can confirm that a function or variable is genuinely unused anywhere in the project before flagging it.
The skill detects unused imports, variables, and private methods; orphaned functions, classes, and exported symbols; unreachable code paths following return or throw statements; leftover debugging statements such as console.log and debugger calls; and large commented-out code blocks.
Language support covers Python, JavaScript, TypeScript, Go, Rust, Java, and C/C++. The analysis is framework-aware: decorators used in Flask or Django and components registered in React or Vue are recognized as active usage, preventing false positives for common entry-point patterns. It also accounts for dynamic usage patterns such as reflection and dependency injection.
The multi-pass workflow validates findings against project-wide searches and handles framework-specific entry points before producing a severity-ranked report. Each finding is classified as either Safe to Remove or Verify Before Removal, giving developers a clear signal about which deletions are low-risk and which require manual review.
This skill is appropriate for codebases being refactored, audited before a release, or inherited by a new team. It is not a runtime profiler and does not detect performance bottlenecks or code that is only unreachable under specific runtime conditions.