feat: switch to nanoclaw image from GHCR

fix: align openclaw.json with correct models schema

fix: move models definition to ai section in config

fix: move doctor fix and add more aggressive feishu bypass

feat: add feishu autoPairing for user ID

fix: dockerfile syntax error and CMD separation

fix: use openai provider for glm model to bypass unknown model error

fix: move custom model definition to agents.models

fix: define custom models in config to resolve Unknown model error

feat: switch model from gemini to anthropic (glm-4.7)

fix: force feishu dmPolicy to open via config and env

feat: use custom gemini base url and set model to gemini-2.5-pro

feat: upgrade primary model to gemini-2.0-pro

feat: set feishu dmPolicy to open to bypass pairing

feat: auto-approve Feishu pairing

fix: change gateway.bind to lan and cleanup startup script

fix: use writable config dir and bind to 0.0.0.0 for HF Spaces

fix: improve startup script robustness and fix config errors

fix: install feishu plugin directly via openclaw command

fix: add curl to system dependencies for plugin installation

feat: integrate Feishu channel with WebSocket mode and local plugin installation

Fix: revert to token auth to bypass device identity checks and allow URL access

Fix: trust all proxies and dangerouslyDisableDeviceAuth for HF Spaces compatibility

Fix: revert allowInsecureAuth to true to ensure UI functionality

Fix: set allowInsecureAuth to false to force password prompt

Fix: correctly configure trustedProxies and extra env vars

Fix: revert to google/gemini-1.5-pro to resolve empty response issue

Fix: switch to env var auth in openclaw.json and remove auth-profiles.json

Fix: rename profile to google-default and add env fallback

Fix: remove stray fi and broken debug command

Fix: force overwrite auth-profiles.json on startup

Debug: verify auth-profiles.json generation

Fix: remove invalid config keys (safetySettings and trustedProxies)

Fix: trust proxies to clean up logs and improve connection handling

Fix: restore auth-profiles.json with correct structure and set model to gemini-2.5-pro

Update default model to google/gemini-2.5-pro

Fix: set Gemini safetySettings to BLOCK_NONE to avoid empty responses

Fix: move API key config to openclaw.json env section

Fix: wrap auth profiles in 'auth' object key

Switch to stable google/gemini-1.5-pro model

Fix: robust auth-profiles.json generation for Gemini

Change gateway auth mode to password using SETUP_PASSWORD

Update default model to google/gemini-2.5-pro

Switch to Gemini model and configure auth-profiles

Fix: revert invalid trustedProxies config to restore service

Fix: enable trustedProxies to silence logs and improve IP detection

Fix: use 'bind': 'lan' which is a valid enum value

Fix: use 'bind': '0.0.0.0' in config instead of invalid CLI flag

Fix: add missing dirs, fix perms, and try --host 0.0.0.0

Debug: enable verbose startup logging to diagnose timeout