context-loader

To load project context efficiently, this Skill uses git for file discovery and respects ignore files. It prioritizes loading manifest, conductor, track, and configuration files to optimize token usage for Conductor workflows.

Token optimization for large project files is achieved by reading only file headers and footers. The loader also enforces a maximum of 15 files per context load to prevent exceeding token limits.

Yes, you can use git ignore files to filter context loading. The loader respects both `.claudeignore` and `.gitignore` files to quickly identify and load only relevant project files for Conductor tracks.

The best way to prepare agents is by loading the prioritized project context. This ensures the agent receives the latest project structure, current track specifications, and task plans before beginning feature implementation.

The limitation for loading project context is a strict maximum of 15 files per context load. Active track specifications and plans are prioritized within this limit to ensure essential information is loaded.

The context loader prioritizes manifest and configuration files to ensure Conductor workflows have the necessary foundational information. This defined loading order prevents redundant work and optimizes token usage.