Code Comment Cleaner is a skill that targets commented-out dead code blocks across eight languages: JavaScript, TypeScript, Python, Java, Go, Rust, HTML, and CSS. It uses a two-pass detection approach. The first pass identifies code indicators inside comments — brackets, keywords, and assignments — to distinguish abandoned logic from genuine prose. The second pass applies preservation rules that explicitly protect TODOs, FIXMEs, developer notes, license headers, copyright notices, disabled tests, and multi-line architectural explanations. Nothing in those categories is touched.
Before any file is modified, the skill generates a preview report listing exactly which lines are targeted and the reason each is flagged. A developer can review this output and confirm before changes are applied. When changes do proceed, the skill writes a .bak backup of the original file first, giving a recovery path if anything looks wrong after the fact.
This is appropriate when a codebase has accumulated commented-out logic over time and manual review would be slow or inconsistent across a large file tree. It handles multi-line comment syntax correctly in all supported languages, which is a common failure point for simple find-and-replace approaches.
It is not a general code formatter, linter, or documentation generator. It does not remove live code, restructure files, or work on languages outside the eight listed. Agents with no tools exposed use this skill’s built-in logic directly rather than through discrete API calls.