RuleChef: Let the LLM Write Rules, Not Answer Every Query

The core mechanism, package structure, and design philosophy are all confirmed accurate against the actual open-source repository, and the technical approach (LLM as rule-compiler, not inference-time dependency) is clearly and specifically explained.

The specific benchmark numbers (78.7 vs 74.8 F1, the feedback-loop improvement figures) are self-reported claims from the same release and haven't been independently verified yet — presented with full confidence as established results rather than "our own measured numbers, pending outside scrutiny."