improve-claude-md

Improve LLM instruction adherence by restructuring CLAUDE.md files with conditional XML scoping, wrapping rules in context-aware blocks that only surface when needed. This systematic file transformation targets software development environments to reduce context noise.

Conditional instruction scoping wraps project rules in XML-based blocks so they only activate when relevant to the current task. This mechanism prevents complex testing patterns from cluttering the context window during routine coding, reducing context noise.

Automate CLAUDE.md refactoring by parsing project history databases to extract recurring feedback and improvement suggestions. Integrating these evolution triggers applies rule-based file transformations to systematically refine project rules.

Optimizing CLAUDE.md requires no external dependencies. The process relies on internal scripts and references to parse project history databases and apply systematic rule-based file transformations for instruction tuning.

The best way to reduce context noise is moving specialized instructions into conditional XML blocks that trigger only during specific tasks. This optimization prevents irrelevant rules from consuming the context window during routine software development.