fix: eliminate gateway dual-start race condition

feat: redirect root / to WebUI chat page

fix: WeChat session persistence - auto-reconnect after gateway restart

fix: bind port 7860 BEFORE backend health checks

fix: SO_REUSEPORT for port binding + error handling in proxy startup

fix: trigger full rebuild

fix: restore multi-line docstring (was broken by single-line edit)

fix: force rebuild to recover from 404 state

fix: detect and auto-restart zombie gateway (executor shutdown)

opt: multi-stage build + auth on diagnostics + connection limiting

fix: robust gateway restart — force-kill residual processes before restart

fix: add Python-based gateway watchdog inside entry.py

feat: add /api/diagnostics/gateway-log endpoint for crash diagnostics

fix: serve WebUI static files directly, proxy /assets/* to client dir

fix: rewrite /webui/* paths and proxy all static assets to BFF

feat: integrate hermes-web-ui BFF server

feat: add pre-built hermes-web-ui (Hermes teal theme + Chinese i18n)

revert: restore pure Python setup (remove Node.js build dependency)

v5.0: Integrate hermes-web-ui (Vue 3 + Naive UI)

add /deploy route - deployment overview page

refactor: remove MemPalace to free ~1-2GB RAM

feat: enable Holographic memory + MemPalace MCP

feat: Qwen3 Coder 480B as primary + Gemma 4 31B as fallback

chore: terminal=cloud, timezone display=北京时间

fix: use PyYAML for config parsing to read nested values correctly

fix: unbuffered logging + ThreadingHTTPServer for real-time SSE

feat: dark-themed real-time dashboard with SSE logs, config panel, interactive controls

fix: correct CLI import (cli->main), add start.sh for persistent symlinks

Hermes Bot: initial deployment with Feishu gateway