What problem does it solve? Code review feedback often lives in chat threads or external tools and gets disconnected from the code it references. This Skill lets you embed structured review annotations directly in source files as comment blocks, then extract them into .aigr sidecar files that a VS Code extension renders as gutter icons, line highlights, and hover tooltips. ## Core Features & Use Cases - Structured Review Blocks: Wrap code in <<<AIGR ... AIGR>>> comment markers with an ID, a [category:LEVEL] label (severity, confidence, or custom), a one-line note, and optional multi-line descriptions. - Validation and Extraction: A bundled Python script validates block formatting (malformed headers, unclosed blocks, missing code, duplicate IDs) and extracts annotations into .aigr JSON sidecars while stripping markers from the source. - Multi-Language Support: Works with Python, JavaScript, TypeScript, C/C++, Rust, Go, Java, Kotlin, C#, Ruby, Shell, YAML, TOML, and more, using each language's native comment syntax. - Use Case: While reviewing a pull request, flag a potential null-pointer bug with a [severity:CRITICAL] block and an N+1 query with a [confidence:MEDIUM] block, then run the extractor so reviewers see the annotations rendered inline in VS Code. ## Quick Start Ask the AI to annotate the current source file with AIGR review blocks flagging potential bugs, then run the bundled extractor script to validate and generate the .aigr sidecar files.
