feat: LLM accuracy testing framework with 17-test catalog

Framework runs a category/subcategory test catalog against multiple
LLMs through an OpenAI-compatible endpoint. Deterministic evaluators
score each response. Reports render as table, markdown, or JSON.

.gitignore

@@ -0,0 +1,12 @@
+# Binaries
+/llmtest
+/bin/
+*.exe
+*.test
+
+# Local config (copy config.example.yaml -> config.yaml and edit)
+/config.yaml
+
+# OS/editor cruft
+.DS_Store
+*.swp