repomix

Package repositories into AI-friendly single-file representations using XML, Markdown, JSON, or plain text format. The skill converts entire codebases—local or remote—into structured files that preserve directory structure, include metadata, and support include/exclude patterns, making code easier for AI agents to process and understand.

Yes, batch-process multiple repositories simultaneously into individual output files. This is ideal for large-scale analysis tasks like security audits across many projects, eliminating the need to clone and process each repository individually while maintaining consistent formatting.

Codebase packaging supports XML, Markdown, JSON, and plain text formats. Each format preserves directory structure and includes per-file token accounting, repository metadata, and configurable headers and summaries tailored to your AI analysis or documentation needs.

Token counting provides per-file accounting to track token usage across packaged repositories. This helps estimate computational costs and manage context limits when sending structured codebase representations to AI models for analysis, audits, or investigations.

Yes, packaging applies to both local repositories and remote GitHub repositories. Git-aware processing handles cloning, configuration management, and security checks automatically, supporting flexible CLI and config-file options for both scenarios.

Security checks are applied during packaging to ensure safe AI consumption. Combined with include/exclude pattern filtering, these checks help prevent sensitive files or configurations from being included in the packaged output for AI analysis.