fix(strategy): tag idle hold state instead of bare wait

fix(strategy): all unit types return to core instead of bare wait

fix(strategy): far vanguard returns to core instead of waiting

fix(strategy): scout retries 8 directions before falling back to wait

chore(memory): refresh space bootstrap from live memory snapshot

fix(strategy): ranger tracking skips enemy memory (use_memory=False)

fix(strategy): ranger ignores enemy workers outside defense

fix(strategy): require dir_streak>=2 before lead-shot prediction

fix(strategy): skip low-hp retreat when Core has no resources

fix(strategy): gate migration on non-idle unit intent only

fix(strategy): pause migration for unit actions

feat(strategy): add station and migration mode toggle

fix(strategy): detour core around obstacles instead of blind retry

feat(web): show stationed and migrating states

fix(strategy): retry automatic migration while empty

fix(strategy): translate frozen migration formation

feat(deploy): persist memory and control core migration

feat(strategy): coordinate whole-population core migration

chore(deploy): promote overnight genes to production

fix(strategy): expand bootstrap scout rings to 10/20/30/40

merge: resolve HF Space remote conflicts and keep early-game patches

feat(strategy): early-game economy boost for live world

Panel: surface upstream SDK fields (Core dest/ETA/state, respawn countdown, unit costs, enemy owner/migration) + failure events

Repurpose Space: arena-evolve (Arena Hero agent + FastAPI panel)

Add strategy reference document (docs/STRATEGY.md)

Add LINUX DO community link to references section

Project documentation, license, and repository hygiene

Live deployment: SDK adapter, deployed genome, dependency manifest

Core research: deterministic simulator, evolutionary strategy, genetic algorithm

Fix actor display and JSON snapshot parsing

Update event summary formatting in memory_helper.py

Rename misleading fallback comment and simplify event summary format

Clean up dead code and eliminate duplication

Refactor accept_deal to accept dealIds list in single call

Update initiate_trade and map_provider

Update game rules, trade tools, map provider and skip tool

Update evaluate_trade handler

Update handlers and tools for trade evaluation

更新各 handler 和 tool_loop

重构 map_sync 工具模块，更新各 handler 和 tools

优化建筑规则提示，修复 map_provider，增强 prompt_builder 和 tool_loop

更新 tool_loop 多模态支持，优化交易相关工具和游戏规则提示

设置 MODEL_NAME 为 gemini-3-flash-preview

Update tools helpers, skip and map provider

Update tools init

Update agent, handlers, prompts, tools and asset manager

refactor(map_provider): 统一棋盘快照解析接口并移除旧视觉解析模块

refactor(tools): 重构工具模块架构以简化维护和扩展实现

Update app and agent code

Add tool-calling support and update handlers