docs: spec for search latency optimization (score gap + LLM fallback)

docs: spec for eval CLI to iterate on retrieval quality

docs: search interface design spec

docs: mosaic redesign spec — masonry, infinite scroll, randomization

docs: custom frontend design spec and implementation plan

docs: design spec for ONNX lightweight migration

spec: corrections review AI engineer (RAG best practices)

spec: intégrer les corrections de la review indépendante

spec: hybrid RAG + enrichissement élargi design

Clarify Gradio rendering and cache compatibility in spec

Add design spec for poster images in RAG pipeline