perf: Apply all thread optimizations + complete documentation

ui: Remove Control Groups section (unnecessary)

feat: Implement cancel-on-new-request strategy (no timeouts)

docs: Explain heuristic vs LLM analysis

refactor: Remove duplicate LLM loading - single shared model only

perf: Non-blocking LLM architecture to prevent game lag

docs: Comprehensive README update with AI Analysis and NL interface features

perf: Optimize LLM for HuggingFace 2 vCPU environment

fix: Change chat_format from 'qwen2' to 'qwen'

fix: Critical fixes for model path and game loop

debug: Add game loop health check logs

debug: Add critical logs to identify movement and production issues

fix: Debug logs and auto-load for NL translator, movement & production bugs

debug: Add extensive logging for unit movement bug investigation

add libopenblas-dev for llama-cpp-python i built with openblas

fix: Shared LLM model manager and UI positioning

update todos

replace qwen2.5 0.5b with qwen2.5 coder 1.5b

Update README.md

Update README.md

use prebuilt llama cpp python for python 3.10, modify docker to use py3.10 instead of py3.11

upgrade llama cpp python to 0.3.16

fix dep conflit 3

fix dependency conflit 2

fix dependency conflit

Add MCP support to the game and integrate NL-to-MCP translation capabilities

Complete project reorganization: Remove moved files from root directory

Organize project structure: move test scripts to tests/scripts and documentation to docs/reports

Initial commit: Complete RTS project with MCP evaluation

fix(ui): prevent camera controls overlap with topbar

fix(ui): prevent minimap overlap with right sidebar

fix(ui): restore minimap base styles

fix(ui): ensure desktop layout works in HF Spaces iframe

fix(ui): correct responsive breakpoints for desktop gaming

feat(ui): modernize UI/UX with new design system

docs: fix paths to moved scripts (tools/ and tests/) in docs

docs: update docs and index to reflect tools/ + moved test scripts

docs(readme): reflect new structure (tools/, docs/, tests/) and update links

chore(structure): add tools/ and move debugging/ops scripts; ignore accidental large artifact

chore(structure): move docs into docs/ and tests into tests/

docs: move todos.txt into web/ (single repo, single branch)

feat(endgame): elimination-based game over + overlay; fix LLM chat_format and i18n draw

deploy(web): full clean snapshot with app code and assets

chore(web): ignore logs and model files; remove server.log from VCS