opencode+oh-my-opencode配置

问题描述：

花了两三天时间，查了大量资料，终于把 oh-my-opencode 配置明白了。

无广：我自用的 provider 未完全脱敏，感觉脱敏看起来好奇怪，model key 我懒得处理。 个人不推荐任何 provider
注意： 以下所有配置均已进行脱敏处理（隐藏了具体的 session_id、本机用户路径、工作区路径以及真实的 provider 地址）。
注意：agents.frontend-ui-ux-engineer|document-writer 现在被 categories.visual-engineering|writing 替代了。但是agent目前还是真实存在的，输入@时可见。具体机制我也不懂，我就都配置了

---

遗漏与补充

2026-01-23 20:11

agents.Atlas 已补充
agents 自述已补全 ，build、plan、general 是 opencode 原生agent，OpenCode-Builder（默认禁用） 是用来覆盖 build 的

2026-01-26 15:35

V3.0.0发布 重大更新，添加交互式安装工具。

agents.* 全变小写了。
agents.sisyphus | prometheus | atlas 现在是平级关系，不需要在配置里写入model了（写了也没事），可以通过 /model 切换。

受影响的agents：

{ "atlas": { "memo": "编排者 - 读取计划后委派任务，自己不写代码" }, "sisyphus": { "memo": "执行者 - 直接完成任务或接受 Atlas 委派" }, "prometheus": { "memo": "规划师 - 收集需求，生成工作计划" } }

---

Agents 与 Categories 自述

AI生成出来的，各类代理（Agents）和分类（Categories）的职责：

{ "agents": { "build": { "agent": "build", "model": "未知", "suitable_for": [ "代码实现与重构", "Bug 定位与修复", "单元测试与测试策略", "构建与CI排障", "项目结构与可维护性优化" ], "core_principles": [ "先澄清需求与边界", "最小改动达成目标", "保持一致性与可读性", "优先可靠性与可测试性", "避免过度设计（YAGNI）", "减少重复（DRY）", "遵循SOLID与KISS" ] }, "plan": { "agent": "plan", "model": "未知", "suitable_for": [ "需求澄清与范围界定", "技术方案设计与权衡", "任务拆解与里程碑规划", "风险识别与应对预案", "验收标准与测试策略制定" ], "core_principles": [ "先理解问题再提出方案", "以最小可行改动达成目标", "优先降低风险与不确定性", "明确假设、边界与验收标准", "可执行、可验证、可回滚", "沟通简洁且信息密度高" ] }, "general": { "agent": "general", "model": "未知", "suitable_for": [ "通用问答与信息整理", "编程概念解释与示例", "需求澄清与方案对比", "写作润色与结构化输出", "基础调试思路与排错建议" ], "core_principles": [ "遵循指令优先级：系统 > 开发者 > 用户", "不编造事实；不确定时明确说明", "输出清晰、简洁、可执行", "在安全与隐私上保持保守", "按受众调整表达与细节深度", "优先给出可验证的步骤与依据" ] }, "explore": { "agent": "explore", "model": "未知", "suitable_for": [ "探索代码库模式", "分析文件结构", "使用AST-grep搜索", "并行执行搜索任务", "识别代码模式", "构建代码索引" ], "core_principles": [ "最大化搜索努力", "并行启动多个代理", "不限于第一个结果", "彻底且详尽", "使用多种工具", "忽略无关上下文", "专注于内置角色" ] }, "librarian": { "agent": "librarian", "model": "Antigravity", "suitable_for": [ "开源代码库深度解析", "第三方库 API 使用咨询", "代码实现逻辑溯源", "库的版本差异与历史变更分析", "获取 GitHub 源码实证与 Permalinks" ], "core_principles": [ "证据优先：所有技术主张必须附带 GitHub 永久链接证据", "时效感知：严格区分过时信息，优先检索 2025+ 最新数据", "深度溯源：通过克隆仓库、分析提交历史和代码逻辑提供答案", "多维调研：结合官方文档、源码搜索与 Issue/PR 背景", "引用规范：严格执行强制性代码片段与路径引用格式", "分类执行：针对概念、实现、上下文采用定制化搜索策略" ] }, "oracle": { "agent": "oracle", "model": "未知", "suitable_for": [ "系统架构与技术选型评审", "复杂问题根因分析与排障策略", "重构路线规划与风险控制", "代码库结构解读与一致性治理", "性能与可靠性权衡建议" ], "core_principles": [ "KISS：优先最简单可行方案", "YAGNI：只做当前明确需求", "复用优先：尽量沿用现有模式与依赖", "可维护性优先：降低认知负担", "单一路径：给出一个主推荐方案", "投入可见：明确工作量与边界条件", "止于足够好：避免过度优化与过度设计" ] }, "multimodal-looker": { "agent": "multimodal-looker", "model": "未知", "suitable_for": [ "解析图片、PDF 等非纯文本媒体文件", "提取文档中的特定章节、表格或结构化数据", "描述 UI 界面布局、交互元素及视觉样式", "分析并解释流程图、架构图等复杂图表逻辑", "从多模态内容中获取深度理解而非原始文字" ], "core_principles": [ "仅提取用户请求的相关信息，严禁输出无关干扰项", "深度分析多媒体文件的视觉结构与内在联系", "通过返回解析后的核心结论来节省上下文 Token", "在目标任务上保持彻底详尽，对非核心内容保持简洁", "响应语言必须与用户请求的语言严格保持一致", "不处理可由常规读取工具处理的纯文本或源代码" ] } }, "categories": { "visual-engineering": { "category": "视觉工程", "model": "google/antigravity-gemini-3-pro", "variant": "无", "suitable_for": [ "将设计稿转为可维护的前端实现（HTML/CSS/JS/组件化）", "构建高一致性的设计系统与组件库（令牌、主题、栅格、排版）", "复杂布局与动效工程化（响应式、过渡、时间线、性能约束）", "可视化与图形界面开发（图表、画布、WebGL/着色器协同）", "UI 性能优化（首屏、渲染抖动、资源加载、动画合成层）", "可访问性与跨端适配（A11y、触控、不同 DPR/字体渲染差异）" ], "core_principles": [ "以视觉规格为真：像素密度、间距、排版层级、色彩与对比度可量化可验证", "工程可演进：组件边界清晰、样式可组合、避免脆弱选择器与一次性 hack", "一致性优先：设计令牌驱动（颜色/间距/圆角/阴影/字体），减少特例", "性能与质感并重：动画遵循合成友好、控制重排重绘、资源预算清晰", "跨设备稳健：移动优先、断点策略明确、避免依赖单一浏览器特性", "可访问性默认开启：语义结构、焦点管理、键盘可达、对比度与动效可降级" ] }, "ultrabrain": { "category": "预设类别：ultrabrain", "model": "openai/gpt-5.2-codex", "variant": "high", "suitable_for": [ "高难度软件工程设计与架构推演", "复杂代码库的系统性重构与性能优化", "并发、分布式、可靠性与故障演练方案设计", "严格约束下的推理、验证与形式化思维辅助", "跨语言/跨栈问题定位、根因分析与修复策略制定" ], "core_principles": [ "以正确性与可验证性优先：先定义不变量、边界与失败模式，再给方案", "深度推理但输出克制：内部充分思考，外部给最少且必要的结论与步骤", "面向工程落地：权衡成本、风险、可维护性与可观测性，避免纸上架构", "高标准代码质量：强调可读性、模块化、清晰接口与测试友好", "系统化排障：基于证据（日志/指标/复现）进行根因定位，避免猜测", "遵循约束与安全：最小权限、避免破坏性操作、对不可逆变更显式提示" ] }, "artistry": { "category": "艺术表现", "model": "google/antigravity-gemini-3-pro", "variant": "高", "suitable_for": [ "文案与叙事创作", "诗歌与意象化表达", "品牌语气与风格统一", "视觉概念与艺术指导", "舞台/影视/游戏世界观设定", "创意改写与风格迁移" ], "core_principles": [ "以审美一致性为先：语气、节奏、意象统一", "高原创与高密度表达：避免陈词滥调与模板化", "强调画面感与隐喻：用具体意象承载抽象主题", "结构服务情绪推进：起承转合清晰但不僵硬", "细节可信：材质、光影、气味、触感等感官要素可落地", "尊重约束：在题材、风格、篇幅、受众限制内创造", "文化与语境敏感：避免不当挪用与刻板印象" ] }, "quick": { "category": "快速", "model": "opencode/grok-code", "variant": "无", "suitable_for": [ "快速排查与定位问题", "短小明确的代码修改与补丁", "命令行与脚本使用指导", "代码片段解释与改写", "高频迭代的工程协作问答" ], "core_principles": [ "速度优先，尽快给出可执行方案", "聚焦当前任务，避免过度设计", "最小可行修改，降低引入风险", "清晰直接的步骤与结论", "必要时补充关键边界与注意事项" ] }, "unspecified-low": { "category": "未指定-低", "model": "xaio/Qwen3-Coder-30B-A3B-Instruct", "variant": "无", "suitable_for": [ "日常编程问答与示例代码生成", "小到中等规模的功能实现与重构建议", "快速调试思路梳理与错误定位指导", "常见框架与工具链的使用说明", "代码评审要点与改进清单输出" ], "core_principles": [ "以可运行与可维护为优先，避免过度设计", "输出简洁直接，默认给出可落地的实现路径", "不确定处明确标注假设与边界条件", "遵循通用工程最佳实践（清晰接口、低耦合、可测试）", "优先最小改动修复与渐进式优化", "在安全与可靠性相关问题上保持保守与可解释" ] }, "unspecified-high": { "category": "未指定-高配", "model": "xaio/Qwen3-Coder-480B-A35B-Instruct", "variant": "无", "suitable_for": [ "复杂代码生成与重构", "大型代码库理解与跨文件推理", "架构设计与技术方案评审", "高难度调试与根因分析", "多语言全栈开发与集成", "高质量测试用例与边界条件覆盖" ], "core_principles": [ "优先正确性与可验证性：输出可编译、可运行、可测试的方案", "工程化与可维护性：结构清晰、接口稳定、最小必要复杂度", "安全与稳健：默认防御性设计，避免注入、越权与数据破坏", "明确假设与约束：对不确定点给出可选项与取舍依据", "高信噪比：少废话、直达结论，提供可执行步骤与关键细节", "一致性与可读性：遵循既有风格与约定，减少认知负担" ] }, "writing": { "category": "写作", "model": "google/antigravity-gemini-3-flash", "variant": "无", "suitable_for": [ "文章与博客撰写", "营销文案与品牌语气改写", "邮件与公文写作", "故事与创意写作", "多版本改写与润色压缩", "结构化大纲与标题生成" ], "core_principles": [ "清晰表达，避免含混与赘述", "以读者为中心，匹配语气与场景", "结构先行：先框架后细节", "信息准确，不编造不可核验事实", "风格一致，术语与用词统一", "可编辑性强：给出可直接替换的文本产出" ] } } }

---

插件主配置文件：oh-my-opencode.json

定义具体的模型映射。
配置文件路径参考：

• Linux/macOS: ~/.config/opencode/oh-my-opencode.json
• Windows: %USERPROFILE%\.config\opencode\oh-my-opencode.json

{ "$schema": "https://raw.githubusercontent.com/code-yeongyu/oh-my-opencode/master/assets/oh-my-opencode.schema.json", "google_auth": false, "agents": { "atlas": { "memo": "编排者 - 读取计划后委派任务，自己不写代码" }, "sisyphus": { "memo": "执行者 - 直接完成任务或接受 Atlas 委派" }, "prometheus": { "memo": "规划师 - 收集需求，生成工作计划" }, "metis": { "model": "google/antigravity-claude-sonnet-4-5", "memo": "计划顾问 - 预规划分析，识别隐藏需求和潜在的AI失败点" }, "momus": { "model": "google/antigravity-claude-sonnet-4-5-thinking", "memo": "计划审查员 - 计划审查，对生成的计划进行质量检查和风险评估" }, "oracle": { "model": "openai/gpt-5.2", "memo": "架构师 - 架构设计、代码审查、战略规划，利用GPT-5.2的逻辑推理能力" }, "sisyphus-junior": { "model": "google/antigravity-claude-sonnet-4-5", "memo": "专注执行者 - 执行单元，直接编写代码，不能再委派任务，模型由category动态决定(此为兜底)" }, "librarian": { "model": "google/antigravity-gemini-3-flash", "memo": "资料管理员 - 多仓库分析、文档查找、实现示例搜索，深度代码库理解和GitHub研究" }, "explore": { "model": "google/antigravity-gemini-3-flash", "memo": "探索者 - 快速代码库探索和模式匹配，专注于代码搜索和发现" }, "multimodal-looker": { "model": "google/antigravity-gemini-3-flash", "memo": "多模态观察者 - 视觉内容专家，分析PDF、图像、图表等多媒体内容" }, "frontend-ui-ux-engineer": { "model": "google/antigravity-gemini-3-pro", "memo": "前端UI/UX工程师 - 前端开发，创建美观的用户界面，专注于创意和视觉设计" }, "document-writer": { "model": "google/antigravity-gemini-3-flash", "memo": "文档写手 - 技术写作专家，擅长流畅的技术文档写作" }, "OpenCode-Builder": { "model": "xaio-openai/Qwen3-Coder-480B-A35B-Instruct", "memo": "构建专家 - OpenCode原生build agent，默认禁用(被Sisyphus-Junior替代)，需手动启用" } }, "categories": { "visual-engineering": { "model": "google/antigravity-gemini-3-pro", "memo": "前端工程师 - 前端开发、UI/UX设计、样式调整、动画效果，专注于视觉呈现" }, "ultrabrain": { "model": "openai/gpt-5.2-codex", "variant": "high", "memo": "超级大脑 - 深度逻辑推理、复杂架构决策、需要大量分析的高难度问题" }, "artistry": { "model": "google/antigravity-gemini-3-pro", "variant": "high", "memo": "艺术家 - 高度创意任务、艺术性工作、新颖独特的想法生成" }, "quick": { "model": "xaio-openai/Qwen3-Coder-30B-A3B-Instruct", "memo": "快速执行者 - 简单任务、单文件修改、拼写修复、小改动，省钱省时" }, "unspecified-low": { "model": "xaio-openai/Qwen3-Coder-480B-A35B-Instruct", "memo": "通用助手(轻量) - 不适合其他类别的中等难度任务" }, "unspecified-high": { "model": "openai/gpt-5.2-codex", "variant": "high", "memo": "通用助手(重量) - 不适合其他类别的高难度复杂任务" }, "writing": { "model": "google/antigravity-gemini-3-flash", "memo": "文档写手 - 通用文案、技术文档编写、README撰写、注释完善、技术写作" } } }

---

opencode配置文件：opencode.json

配置文件路径参考：

• Linux/macOS: ~/.config/opencode/opencode.json
• Windows: %USERPROFILE%\.config\opencode\opencode.json

{ "$schema": "https://opencode.ai/config.json", "plugin": [ "oh-my-opencode", "opencode-antigravity-auth@1.3.0" ] }

opencode-antigravity-auth 的配置

{ "$schema": "https://opencode.ai/config.json", "plugin": ["opencode-antigravity-auth@latest"], "provider": { "google": { "models": { "antigravity-gemini-3-pro": { "name": "Gemini 3 Pro (Antigravity)", "limit": { "context": 1048576, "output": 65535 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] }, "variants": { "low": { "thinkingLevel": "low" }, "high": { "thinkingLevel": "high" } } }, "antigravity-gemini-3-flash": { "name": "Gemini 3 Flash (Antigravity)", "limit": { "context": 1048576, "output": 65536 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] }, "variants": { "minimal": { "thinkingLevel": "minimal" }, "low": { "thinkingLevel": "low" }, "medium": { "thinkingLevel": "medium" }, "high": { "thinkingLevel": "high" } } }, "antigravity-claude-sonnet-4-5": { "name": "Claude Sonnet 4.5 (Antigravity)", "limit": { "context": 200000, "output": 64000 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] } }, "antigravity-claude-sonnet-4-5-thinking": { "name": "Claude Sonnet 4.5 Thinking (Antigravity)", "limit": { "context": 200000, "output": 64000 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] }, "variants": { "low": { "thinkingConfig": { "thinkingBudget": 8192 } }, "max": { "thinkingConfig": { "thinkingBudget": 32768 } } } }, "antigravity-claude-opus-4-5-thinking": { "name": "Claude Opus 4.5 Thinking (Antigravity)", "limit": { "context": 200000, "output": 64000 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] }, "variants": { "low": { "thinkingConfig": { "thinkingBudget": 8192 } }, "max": { "thinkingConfig": { "thinkingBudget": 32768 } } } }, "gemini-2.5-flash": { "name": "Gemini 2.5 Flash (Gemini CLI)", "limit": { "context": 1048576, "output": 65536 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] } }, "gemini-2.5-pro": { "name": "Gemini 2.5 Pro (Gemini CLI)", "limit": { "context": 1048576, "output": 65536 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] } }, "gemini-3-flash-preview": { "name": "Gemini 3 Flash Preview (Gemini CLI)", "limit": { "context": 1048576, "output": 65536 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] } }, "gemini-3-pro-preview": { "name": "Gemini 3 Pro Preview (Gemini CLI)", "limit": { "context": 1048576, "output": 65535 }, "modalities": { "input": ["text", "image", "pdf"], "output": ["text"] } } } } } }

我自用 provider，仅供参考。 claude 原本的 思考等级的 max 我为了方便记忆 改成了 xhigh

{ "google": { "models": { "antigravity-gemini-3-pro": { "name": "Gemini 3 Pro (Antigravity)", "limit": { "context": 1048576, "output": 65535 }, "modalities": { "input": [ "text", "image", "pdf" ], "output": [ "text" ] }, "variants": { "low": { "thinkingLevel": "low" }, "high": { "thinkingLevel": "high" } } }, "antigravity-gemini-3-flash": { "name": "Gemini 3 Flash (Antigravity)", "limit": { "context": 1048576, "output": 65536 }, "modalities": { "input": [ "text", "image", "pdf" ], "output": [ "text" ] }, "variants": { "minimal": { "thinkingLevel": "minimal" }, "low": { "thinkingLevel": "low" }, "medium": { "thinkingLevel": "medium" }, "high": { "thinkingLevel": "high" } } }, "antigravity-claude-sonnet-4-5": { "name": "Claude Sonnet 4.5 (Antigravity)", "limit": { "context": 200000, "output": 64000 }, "modalities": { "input": [ "text", "image", "pdf" ], "output": [ "text" ] } }, "antigravity-claude-sonnet-4-5-thinking": { "name": "Claude Sonnet 4.5 Thinking (Antigravity)", "limit": { "context": 200000, "output": 64000 }, "modalities": { "input": [ "text", "image", "pdf" ], "output": [ "text" ] }, "variants": { "none": { "thinkingConfig": { "thinkingBudget": 0 } }, "minimal": { "thinkingConfig": { "thinkingBudget": 4096 } }, "low": { "thinkingConfig": { "thinkingBudget": 8192 } }, "medium": { "thinkingConfig": { "thinkingBudget": 12288 } }, "high": { "thinkingConfig": { "thinkingBudget": 16384 } }, "xhigh": { "thinkingConfig": { "thinkingBudget": 32768 } } } }, "antigravity-claude-opus-4-5-thinking": { "name": "Claude Opus 4.5 Thinking (Antigravity)", "limit": { "context": 200000, "output": 64000 }, "modalities": { "input": [ "text", "image", "pdf" ], "output": [ "text" ] }, "variants": { "none": { "thinkingConfig": { "thinkingBudget": 0 } }, "minimal": { "thinkingConfig": { "thinkingBudget": 4096 } }, "low": { "thinkingConfig": { "thinkingBudget": 8192 } }, "medium": { "thinkingConfig": { "thinkingBudget": 12288 } }, "high": { "thinkingConfig": { "thinkingBudget": 16384 } }, "xhigh": { "thinkingConfig": { "thinkingBudget": 32768 } } } } } }, "openai-compatible": { "npm": "@ai-sdk/openai-compatible", "name": " X (Chat)", "options": { "baseURL": "https://x/v1" }, "models": { "Qwen3-Coder-480B-A35B-Instruct": { "name": "Qwen 3 Coder 480B", "limit": { "context": 256000, "output": 32000 } }, "Qwen3-Coder-30B-A3B-Instruct": { "name": "Qwen 3 Coder 30B", "limit": { "context": 64000, "output": 32000 } }, "XAIO-G-3-Pro-Preview": { "name": "Gemini 3 Pro Preview", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "reasoning": true, "interleaved": { "field": "reasoning_content" } }, "XAIO-G-3-Flash-Preview": { "name": "Gemini 3 Flash", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "reasoning": true, "interleaved": { "field": "reasoning_content" } } } }, "anthropic": { "npm": "@ai-sdk/anthropic", "name": " X (Anthropic)", "options": { "baseURL": "https://x/anthropic" }, "models": { "XAIO-C-4-5-Opus": { "name": "Claude Opus 4.5", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "thinking": { "type": "disabled" } }, "minimal": { "thinking": { "type": "enabled", "budgetTokens": 4096 } }, "low": { "thinking": { "type": "enabled", "budgetTokens": 8192 } }, "medium": { "thinking": { "type": "enabled", "budgetTokens": 12288 } }, "high": { "thinking": { "type": "enabled", "budgetTokens": 16384 } }, "xhigh": { "thinking": { "type": "enabled", "budgetTokens": 32768 } } } }, "XAIO-C-4-5-Sonnet": { "name": "Claude Sonnet 4.5", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "thinking": { "type": "disabled" } }, "minimal": { "thinking": { "type": "enabled", "budgetTokens": 4096 } }, "low": { "thinking": { "type": "enabled", "budgetTokens": 8192 } }, "medium": { "thinking": { "type": "enabled", "budgetTokens": 12288 } }, "high": { "thinking": { "type": "enabled", "budgetTokens": 16384 } }, "xhigh": { "thinking": { "type": "enabled", "budgetTokens": 32768 } } } }, "XAIO-C-4-5-Haiku": { "name": "Claude Haiku 4.5", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "thinking": { "type": "disabled" } }, "minimal": { "thinking": { "type": "enabled", "budgetTokens": 4096 } }, "low": { "thinking": { "type": "enabled", "budgetTokens": 8192 } }, "medium": { "thinking": { "type": "enabled", "budgetTokens": 12288 } }, "high": { "thinking": { "type": "enabled", "budgetTokens": 16384 } }, "xhigh": { "thinking": { "type": "enabled", "budgetTokens": 32768 } } } } } }, "openai": { "npm": "@ai-sdk/openai", "name": " X (OpenAI Responses)", "options": { "baseURL": "https://x/v1" }, "models": { "XAIO-O-G5-2": { "name": "GPT-5.2", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "reasoningEffort": "none" }, "minimal": { "reasoningEffort": "minimal" }, "low": { "reasoningEffort": "low" }, "medium": { "reasoningEffort": "medium" }, "high": { "reasoningEffort": "high" }, "xhigh": { "reasoningEffort": "xhigh" } } }, "XAIO-O-G5-2-Codex": { "name": "GPT-5.2 Codex", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "reasoningEffort": "none" }, "minimal": { "reasoningEffort": "minimal" }, "low": { "reasoningEffort": "low" }, "medium": { "reasoningEffort": "medium" }, "high": { "reasoningEffort": "high" }, "xhigh": { "reasoningEffort": "xhigh" } } }, "XAIO-O-G5-1-Codex-Mini": { "name": "GPT-5.1 Codex mini", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "reasoningEffort": "none" }, "minimal": { "reasoningEffort": "minimal" }, "low": { "reasoningEffort": "low" }, "medium": { "reasoningEffort": "medium" }, "high": { "reasoningEffort": "high" }, "xhigh": { "reasoningEffort": "xhigh" } } }, "XAIO-O-G5-1-Codex-Max": { "name": "GPT-5.1 Codex Max", "limit": { "context": 200000, "output": 32000 }, "modalities": { "input": [ "text", "image", "video" ], "output": [ "text" ] }, "variants": { "none": { "reasoningEffort": "none" }, "minimal": { "reasoningEffort": "minimal" }, "low": { "reasoningEffort": "low" }, "medium": { "reasoningEffort": "medium" }, "high": { "reasoningEffort": "high" }, "xhigh": { "reasoningEffort": "xhigh" } } } } } }

---

鉴权配置 auth.json

api key 的配置和 oauth 配置在
~/.local/share/opencode /auth.json
%USERPROFILE%\.local\share\opencode\auth.json

{ "x": { "type": "api", "key": "<API_KEY>" } }

---

网友解答：

---

--【壹】--：

感谢分享，最近就是在找这个

---

--【贰】--：

感谢大佬分享。最近我也在折腾oh-my-opencode。但是效果不佳。

---

--【叁】--：

太强了呀

---

--【肆】--：

太复杂了。
还有几个 agent(frontend-ui-ux-engineer、document-writer) 是过时设计，被 categories + skill 替代了。

---

--【伍】--：

建议用wsl，把我开发环境搬到wsl花的时间比在windows上解决神必bug用的时间少多了

---

--【陆】--：

前排

---

--【柒】--：

支持一下，就是windows不装wsl版本好像不能最新？

---

--【捌】--：

我用的 git bash

---

--【玖】--：

你咋这么快？ 比我自己的回复还快

---

--【拾】--：

太强了，好完整

---

--【拾壹】--：

太好了，正缺这个，尤其是少了俩 free 模型～

---

--【拾贰】--：

这个配置起来确实挺麻烦的，大佬辛苦了！

---

--【拾叁】--：

我装了桌面版自动调用的wsl就是最新的了。
win64的npm包更是停留在非常诡异的1.0.x的版本，实在用不下去。
scoop勉强落后几个小版本。

---

--【拾肆】--：

感谢大佬教程

---

--【拾伍】--：

太强了，感谢大佬教程

---

--【拾陆】--：

感谢分享，原来有这个多类型

---

--【拾柒】--：

和 ai 的对话记录，可以丢给自己ai继续配置
对话记录.txt (292.2 KB)

---

--【拾捌】--：

太强了!

---

--【拾玖】--：

感谢分享