COMMITS
July 16, 2026
V
fix(commit): 移除提交消息生成成功时的信息提示弹窗 #303
Vicanent committed
V
feat(streamlake): 调整 KAT-Coder 系列模型 token 限制参数 #301
Vicanent committed
V
refactor(docs): 统一 JSON 配置格式，移除冗余示例代码
Vicanent committed
July 15, 2026
V
feat(retry): 优化 provider 级重试配置合并逻辑，区分 preset 与 override 语义
Vicanent committed
V
refactor(anthropic): 提取 thinking/output_config 配置逻辑为独立模块 #300
Vicanent committed
V
feat(retry): 新增重试状态栏提示与进度包装机制
Vicanent committed
V
feat: 新增提供商级别重试配置覆盖，支持无限重试与子 provider 独立策略
Vicanent committed
V
feat: 新增 editTools 能力声明支持，允许配置模型偏好的编辑工具
Vicanent committed
Z
feat: 新增 editTools 能力声明字段并应用于 GLM-5.2 (#299)
zhangxp committed
V
0.25.37 #298
Vicanent committed
July 14, 2026
V
feat: 新增百度千帆 Token Plan 个人版与企业版接入支持 #293
Vicanent committed
V
feat(streamlake): 新增 KAT-Coder V2.5 系列模型支持 #296
Vicanent committed
July 13, 2026
V
0.25.35 #279
Vicanent committed
V
0.25.34 #279
Vicanent committed
V
0.25.33 #279
Vicanent committed
July 11, 2026
July 10, 2026
V
0.25.31 #277
Vicanent committed
V
refactor: 重构 Codex 模型解析逻辑，简化远端与本地模型合并策略
Vicanent committed
叫
feat: 新增Codex自动从后端拉取可用模型 (#288)
叫我沈同学 committed
V
0.25.30 #287
Vicanent committed
V
feat: FIM/NES 新增熔断器机制防止连续失败时无限重试 (#279)
Vicanent committed
July 9, 2026
V
0.25.28 #283
Vicanent committed
V
0.25.27 #284
Vicanent committed