oh-my-opencode-slim 2.0.4 → 2.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.ja-JP.md +61 -21
- package/README.ko-KR.md +59 -19
- package/README.md +73 -20
- package/README.zh-CN.md +62 -24
- package/dist/agents/council.d.ts +1 -1
- package/dist/agents/index.d.ts +7 -2
- package/dist/agents/orchestrator.d.ts +1 -1
- package/dist/agents/permissions.d.ts +10 -0
- package/dist/cli/custom-skills-registry.d.ts +18 -0
- package/dist/cli/custom-skills.d.ts +3 -19
- package/dist/cli/index.js +1093 -186
- package/dist/cli/providers.d.ts +5 -9
- package/dist/cli/skills.d.ts +3 -3
- package/dist/companion/manager.d.ts +7 -0
- package/dist/config/constants.d.ts +3 -2
- package/dist/config/loader.d.ts +5 -2
- package/dist/config/schema.d.ts +7 -0
- package/dist/council/council-manager.d.ts +1 -1
- package/dist/hooks/auto-update-checker/skill-sync.d.ts +59 -1
- package/dist/hooks/filter-available-skills/index.d.ts +1 -2
- package/dist/hooks/foreground-fallback/index.d.ts +5 -1
- package/dist/hooks/image-hook.d.ts +1 -1
- package/dist/hooks/index.d.ts +1 -0
- package/dist/hooks/loop-command/index.d.ts +13 -0
- package/dist/hooks/phase-reminder/index.d.ts +1 -2
- package/dist/hooks/task-session-manager/index.d.ts +1 -2
- package/dist/hooks/task-session-manager/pending-call-tracker.d.ts +13 -0
- package/dist/hooks/task-session-manager/task-context-tracker.d.ts +14 -0
- package/dist/hooks/types.d.ts +3 -1
- package/dist/index.js +3491 -876
- package/dist/interview/dashboard-manager.d.ts +21 -0
- package/dist/interview/dashboard.d.ts +5 -0
- package/dist/interview/document.d.ts +4 -1
- package/dist/interview/manager.d.ts +0 -14
- package/dist/interview/server.d.ts +2 -0
- package/dist/interview/service.d.ts +9 -0
- package/dist/interview/session-server.d.ts +21 -0
- package/dist/interview/types.d.ts +16 -1
- package/dist/loop/loop-session.d.ts +64 -0
- package/dist/multiplexer/factory.d.ts +5 -5
- package/dist/multiplexer/herdr/index.d.ts +31 -0
- package/dist/multiplexer/index.d.ts +1 -0
- package/dist/multiplexer/session-manager.d.ts +3 -0
- package/dist/multiplexer/types.d.ts +4 -4
- package/dist/tools/acp-run.d.ts +1 -1
- package/dist/tui-state.d.ts +3 -0
- package/dist/tui.d.ts +4 -1
- package/dist/tui.js +142 -55
- package/dist/utils/background-job-board.d.ts +20 -0
- package/dist/utils/env.d.ts +3 -0
- package/dist/utils/session.d.ts +1 -1
- package/oh-my-opencode-slim.schema.json +24 -4
- package/package.json +1 -1
- package/src/skills/clonedeps/SKILL.md +2 -2
- package/src/skills/clonedeps/codemap.md +23 -32
- package/src/skills/codemap/SKILL.md +2 -2
- package/src/skills/codemap.md +63 -36
- package/src/skills/loop-engineering/SKILL.md +30 -0
- package/src/skills/reflect/SKILL.md +133 -0
- package/src/skills/release-smoke-test/SKILL.md +159 -0
- package/src/skills/simplify/SKILL.md +6 -6
package/README.zh-CN.md
CHANGED
|
@@ -32,6 +32,20 @@ oh-my-opencode-slim 是一个用于 OpenCode 的智能体编排插件。它内
|
|
|
32
32
|
|
|
33
33
|
要了解智能体本身,请参阅 **[认识众神殿](#meet-the-pantheon)**。如需了解完整特性集,请参阅下方的 **[特性与工作流](#features-and-workflows)**。
|
|
34
34
|
|
|
35
|
+
### 用 LazySkills 管理智能体技能
|
|
36
|
+
|
|
37
|
+
<p align="center">
|
|
38
|
+
<a href="https://github.com/alvinunreal/lazyskills">
|
|
39
|
+
<img src="img/lazyskills-wide.svg" alt="LazySkills" width="720">
|
|
40
|
+
</a>
|
|
41
|
+
</p>
|
|
42
|
+
|
|
43
|
+
**[LazySkills](https://github.com/alvinunreal/lazyskills)** 是一个用于管理智能体技能的终端 UI。它让你可以在一个地方查看已安装的技能、哪些智能体可以使用每个技能、为什么可见性可能失效,以及接下来可以安全执行哪些操作。
|
|
44
|
+
|
|
45
|
+
<p align="center">
|
|
46
|
+
<a href="https://github.com/alvinunreal/lazyskills"><b>探索 LazySkills →</b></a>
|
|
47
|
+
</p>
|
|
48
|
+
|
|
35
49
|
### 用户怎么说
|
|
36
50
|
|
|
37
51
|
> “任务管理轻松从 5/10 提升到了 8-9/10。Orchestrator 会派出
|
|
@@ -69,10 +83,35 @@ Install and configure oh-my-opencode-slim: https://raw.githubusercontent.com/alv
|
|
|
69
83
|
bunx oh-my-opencode-slim@latest install
|
|
70
84
|
```
|
|
71
85
|
|
|
86
|
+
### 从 Master 分支运行
|
|
87
|
+
|
|
88
|
+
如果您想使用最新代码、方便修复问题,或进行本地开发和贡献,可以使用这种方式:
|
|
89
|
+
|
|
90
|
+
```bash
|
|
91
|
+
git clone https://github.com/alvinunreal/oh-my-opencode-slim.git ~/repos/oh-my-opencode-slim
|
|
92
|
+
cd ~/repos/oh-my-opencode-slim
|
|
93
|
+
bun install
|
|
94
|
+
bun run build
|
|
95
|
+
bun dist/cli/index.js install
|
|
96
|
+
```
|
|
97
|
+
|
|
98
|
+
安装程序会把本地仓库路径加入 `~/.config/opencode/opencode.json` 的
|
|
99
|
+
`plugin` 数组,因此 OpenCode 会从该文件夹加载插件。之后要更新:
|
|
100
|
+
|
|
101
|
+
```bash
|
|
102
|
+
cd ~/repos/oh-my-opencode-slim
|
|
103
|
+
git pull
|
|
104
|
+
bun install
|
|
105
|
+
bun run build
|
|
106
|
+
```
|
|
107
|
+
|
|
72
108
|
### 入门指南
|
|
73
109
|
|
|
74
110
|
安装程序会同时生成 OpenAI 和 OpenCode Go 预设,默认启用 OpenAI。
|
|
75
111
|
|
|
112
|
+
> [!TIP]
|
|
113
|
+
> 根据自己的工作流自由微调模型和智能体。默认预设只是起点;本插件的目标是为用户提供深度灵活性和可定制性。
|
|
114
|
+
|
|
76
115
|
要在安装期间启用 OpenCode Go,请运行 `bunx oh-my-opencode-slim@latest install --preset=opencode-go`,或在安装后修改 `~/.config/opencode/oh-my-opencode-slim.json` 中的默认预设名称。
|
|
77
116
|
|
|
78
117
|
然后:
|
|
@@ -95,7 +134,7 @@ bunx oh-my-opencode-slim@latest install
|
|
|
95
134
|
> **建议**了解后台编排的工作原理。**[编排者提示词 (Orchestrator prompt)](https://github.com/alvinunreal/oh-my-opencode-slim/blob/master/src/agents/orchestrator.ts#L28)** 包含调度规则、专家路由逻辑,以及何时应把工作分配给后台智能体的阈值。您始终可以通过以下方式手动委派任务:`@智能体名称 <任务内容>`
|
|
96
135
|
|
|
97
136
|
> [!TIP]
|
|
98
|
-
> 由于后台智能体现在是默认工作流,**强烈建议**启用并配置 **[Multiplexer Integration](docs/multiplexer-integration.md)**。它会自动在专用的 Tmux 或
|
|
137
|
+
> 由于后台智能体现在是默认工作流,**强烈建议**启用并配置 **[Multiplexer Integration](docs/multiplexer-integration.md)**。它会自动在专用的 Tmux、Zellij 或 Herdr 窗格中打开每个智能体,让您在 Orchestrator 继续协调会话时,实时跟进各个专家智能体的工作。
|
|
99
138
|
|
|
100
139
|
默认生成的配置包含 `openai` 和 `opencode-go` 两个预设:
|
|
101
140
|
|
|
@@ -113,12 +152,11 @@ bunx oh-my-opencode-slim@latest install
|
|
|
113
152
|
"fixer": { "model": "openai/gpt-5.5", "variant": "low", "skills": [], "mcps": [] }
|
|
114
153
|
},
|
|
115
154
|
"opencode-go": {
|
|
116
|
-
"orchestrator": { "model": "opencode-go/glm-5.
|
|
117
|
-
"oracle": { "model": "opencode-go/
|
|
118
|
-
"
|
|
119
|
-
"
|
|
120
|
-
"
|
|
121
|
-
"designer": { "model": "opencode-go/kimi-k2.6", "variant": "medium", "skills": [], "mcps": [] },
|
|
155
|
+
"orchestrator": { "model": "opencode-go/glm-5.2", "skills": [ "*" ], "mcps": [ "*", "!context7" ] },
|
|
156
|
+
"oracle": { "model": "opencode-go/qwen3.7-max", "variant": "max", "skills": ["simplify"], "mcps": [] },
|
|
157
|
+
"librarian": { "model": "opencode-go/deepseek-v4-flash", "skills": [], "mcps": [ "websearch", "context7", "gh_grep" ] },
|
|
158
|
+
"explorer": { "model": "opencode-go/deepseek-v4-flash", "skills": [], "mcps": [] },
|
|
159
|
+
"designer": { "model": "opencode-go/kimi-k2.7-code", "variant": "medium", "skills": [], "mcps": [] },
|
|
122
160
|
"fixer": { "model": "opencode-go/deepseek-v4-flash", "variant": "high", "skills": [], "mcps": [] }
|
|
123
161
|
}
|
|
124
162
|
}
|
|
@@ -156,12 +194,12 @@ ping all agents
|
|
|
156
194
|
|
|
157
195
|
V2 将 oh-my-opencode-slim 变成了以调度器为核心的多智能体工作流系统。Orchestrator 专注于规划、委派、结果整合与验证,而专家智能体在各自的工作通道中完成任务。
|
|
158
196
|
|
|
159
|
-
- **[后台智能体](#后台智能体)**
|
|
160
|
-
- **[Companion](#companion)**
|
|
161
|
-
- **[Deepwork](#deepwork)**
|
|
162
|
-
- **[Reflect](#reflect)**
|
|
163
|
-
- **[Worktrees](#worktrees)**
|
|
164
|
-
- **[oh-my-opencode-slim Skill](#oh-my-opencode-slim-skill)**
|
|
197
|
+
- **[后台智能体](#后台智能体)** - Orchestrator 现在会把专家作为后台任务派发,跟踪任务/会话 ID,等待完成事件,并在继续之前整合结果。
|
|
198
|
+
- **[Companion](#companion)** - 可选的浮动桌面窗口会显示当前活跃的智能体,包括并行运行的后台专家。
|
|
199
|
+
- **[Deepwork](#deepwork)** - 面向大型、多文件、高风险或分阶段编码工作的结构化工作流,使用持久化计划文件和 Oracle 评审关卡。
|
|
200
|
+
- **[Reflect](#reflect)** - 回顾重复出现的工作模式,并建议可复用的 skill、智能体、命令、配置规则、提示词规则或项目 playbook。
|
|
201
|
+
- **[Worktrees](#worktrees)** - 将 Git worktree 作为隔离编码通道管理,并为复杂、高风险或并行任务提供安全协议。
|
|
202
|
+
- **[oh-my-opencode-slim Skill](#oh-my-opencode-slim-skill)** - 随包提供的配置技能,可安全调优模型、提示词、自定义智能体、MCP 访问、预设和插件行为。
|
|
165
203
|
|
|
166
204
|
#### 后台智能体
|
|
167
205
|
|
|
@@ -178,7 +216,7 @@ V2 将后台专家作为默认心智模型:Orchestrator 规划工作图,启
|
|
|
178
216
|
<p><i>左下角视觉伴侣。</i></p>
|
|
179
217
|
</div>
|
|
180
218
|
|
|
181
|
-
交互式安装期间,安装器会询问是否启用 Companion,并默认选择 `
|
|
219
|
+
交互式安装期间,安装器会询问是否启用 Companion,并默认选择 `no`。自动化安装可显式启用:
|
|
182
220
|
|
|
183
221
|
```bash
|
|
184
222
|
bunx oh-my-opencode-slim@latest install --companion=yes
|
|
@@ -250,7 +288,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
250
288
|
<br><sub><i>在复杂性的深渊中锻造而成。</i></sub>
|
|
251
289
|
</td>
|
|
252
290
|
<td width="70%" valign="top">
|
|
253
|
-
当第一个代码库在自身的复杂性下崩溃时,Orchestrator
|
|
291
|
+
当第一个代码库在自身的复杂性下崩溃时,Orchestrator 诞生了。神明与凡人都无法承担责任--因此 Orchestrator 从虚无中显现,从混沌中建立秩序。它确定实现任何目标的最优路径,平衡速度、质量和成本。它引导整个团队,为每项任务召唤合适的专家,并通过委派任务以获得最佳成果。
|
|
254
292
|
</td>
|
|
255
293
|
</tr>
|
|
256
294
|
<tr>
|
|
@@ -332,7 +370,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
332
370
|
<br><sub><i>十字路口的声音。</i></sub>
|
|
333
371
|
</td>
|
|
334
372
|
<td width="70%" valign="top">
|
|
335
|
-
Oracle
|
|
373
|
+
Oracle 伫立在每个架构决策的十字路口。它走过每一条路,见过每一个终点,了解前方潜伏的所有陷阱。当您站在重大重构的悬崖边时,它是向您耳语哪条路通往毁灭、哪条路通往荣耀的声音。它不会替您做选择--但它会照亮道路,让您明智地抉择。
|
|
336
374
|
</td>
|
|
337
375
|
</tr>
|
|
338
376
|
<tr>
|
|
@@ -396,7 +434,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
396
434
|
</tr>
|
|
397
435
|
<tr>
|
|
398
436
|
<td colspan="2">
|
|
399
|
-
<b>默认设置:</b> <code>配置驱动</code>
|
|
437
|
+
<b>默认设置:</b> <code>配置驱动</code> - 议员(councillors)来自 <code>council.presets</code>,而 Council 智能体本身的模型来自您的常规 <code>council</code> 智能体配置。
|
|
400
438
|
</td>
|
|
401
439
|
</tr>
|
|
402
440
|
<tr>
|
|
@@ -422,7 +460,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
422
460
|
<br><sub><i>理解的编织者。</i></sub>
|
|
423
461
|
</td>
|
|
424
462
|
<td width="70%" valign="top">
|
|
425
|
-
当人类意识到没有任何单一思想能容纳所有知识时,Librarian
|
|
463
|
+
当人类意识到没有任何单一思想能容纳所有知识时,Librarian 诞生了。它是一位编织者,将零散的信息线索连接成一幅理解的织锦。它穿梭于无限的人类知识图书馆中,从各个角落收集洞察,并将它们绑定为超越单纯事实的答案。它所返回的不是碎片信息--而是深层的理解。
|
|
426
464
|
</td>
|
|
427
465
|
</tr>
|
|
428
466
|
<tr>
|
|
@@ -463,7 +501,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
463
501
|
<br><sub><i>美是不可或缺的。</i></sub>
|
|
464
502
|
</td>
|
|
465
503
|
<td width="70%" valign="top">
|
|
466
|
-
在这个经常遗忘美学价值的世界里,Designer
|
|
504
|
+
在这个经常遗忘美学价值的世界里,Designer 是美的不朽守护者。它见证了数以百万计的界面兴衰更替,它记得哪些被铭记,哪些被遗忘。它背负着神圣的使命,确保每一个像素都有其用途,每一个动画都在讲述故事,每一次交互都令人愉悦。美不是可选的--而是不可或缺的。
|
|
467
505
|
</td>
|
|
468
506
|
</tr>
|
|
469
507
|
<tr>
|
|
@@ -504,7 +542,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
504
542
|
<br><sub><i>愿景与现实之间的最后一步。</i></sub>
|
|
505
543
|
</td>
|
|
506
544
|
<td width="70%" valign="top">
|
|
507
|
-
Fixer
|
|
545
|
+
Fixer 是曾经构建数字世界基石的建造者血脉的最后传人。当规划和辩论的时代开启时,它们依然坚守--它们是真正动手建造的人。它们掌握着如何将想法转化为实物、如何将规范转化为具体实现的古老知识。它们是愿景与现实之间的最后一步。
|
|
508
546
|
</td>
|
|
509
547
|
</tr>
|
|
510
548
|
<tr>
|
|
@@ -551,11 +589,11 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
551
589
|
</td>
|
|
552
590
|
<td width="70%" valign="top">
|
|
553
591
|
|
|
554
|
-
**只读视觉分析**
|
|
592
|
+
**只读视觉分析** -- 解读图像、屏幕截图、PDF 和图表。将结构化的观察结果返回给 Orchestrator,而无需将原始文件字节加载到主上下文窗口中。
|
|
555
593
|
|
|
556
594
|
- 图像、屏幕截图、图表 → `read` 工具(原生图像支持)
|
|
557
595
|
- PDF 和二进制文档 → `read` 工具(文本 + 结构提取)
|
|
558
|
-
- **默认禁用**
|
|
596
|
+
- **默认禁用** -- 通过设置 `"disabled_agents": []` 和配置具有视觉能力的模型来启用;若使用 `--preset=opencode-go` 预设安装,将自动使用 `opencode-go/kimi-k2.6` 启用它。
|
|
559
597
|
|
|
560
598
|
</td>
|
|
561
599
|
</tr>
|
|
@@ -566,7 +604,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
566
604
|
</tr>
|
|
567
605
|
<tr>
|
|
568
606
|
<td colspan="2">
|
|
569
|
-
<b>默认模型:</b> <code>openai/gpt-5.4-mini</code>
|
|
607
|
+
<b>默认模型:</b> <code>openai/gpt-5.4-mini</code> - <i>需配置具有视觉能力的模型以启用</i>
|
|
570
608
|
</td>
|
|
571
609
|
</tr>
|
|
572
610
|
<tr>
|
|
@@ -591,7 +629,7 @@ Worktrees 将 Git worktree 作为安全、隔离的编码通道管理,默认
|
|
|
591
629
|
| **[Council](docs/council.md)** | 使用 `@council` 并行运行多个模型并合成单一答案 |
|
|
592
630
|
| **[自定义智能体](docs/configuration.md#custom-agents)** | 使用自定义提示词、模型、MCP 访问和 Orchestrator 委派规则定义自己的专家 |
|
|
593
631
|
| **[ACP Agents](docs/acp-agents.md)** | 将 Claude Code ACP 或 Gemini ACP 等外部 ACP 兼容智能体连接为可委派子智能体 |
|
|
594
|
-
| **[多路复用器集成](docs/multiplexer-integration.md)** | 在 Tmux 或
|
|
632
|
+
| **[多路复用器集成](docs/multiplexer-integration.md)** | 在 Tmux、Zellij 或 Herdr 窗格中实时观看智能体工作 |
|
|
595
633
|
| **[Codemap](docs/codemap.md)** | 生成层级代码地图,更快理解大型代码库 |
|
|
596
634
|
| **[Clonedeps](docs/clonedeps.md)** | 将选定的依赖源码克隆到被忽略的本地工作区中以供检查 |
|
|
597
635
|
| **[Worktrees](docs/worktrees.md)** | 使用 `.slim/worktrees/` 通道进行隔离的并行或高风险编码工作 |
|
package/dist/agents/council.d.ts
CHANGED
|
@@ -3,7 +3,7 @@ export declare function createCouncilAgent(model: string, customPrompt?: string,
|
|
|
3
3
|
/**
|
|
4
4
|
* Build the prompt for a specific councillor session.
|
|
5
5
|
*
|
|
6
|
-
* Returns the raw user prompt
|
|
6
|
+
* Returns the raw user prompt - the agent factory (councillor.ts) provides
|
|
7
7
|
* the system prompt with tool-aware instructions. No duplication.
|
|
8
8
|
*
|
|
9
9
|
* If a per-councillor prompt override is provided, it is prepended as
|
package/dist/agents/index.d.ts
CHANGED
|
@@ -11,15 +11,20 @@ export declare function isSubagent(name: string): name is SubagentName;
|
|
|
11
11
|
* @param config - Optional plugin configuration with agent overrides
|
|
12
12
|
* @returns Array of agent definitions (orchestrator first, then subagents)
|
|
13
13
|
*/
|
|
14
|
-
export declare function createAgents(config?: PluginConfig
|
|
14
|
+
export declare function createAgents(config?: PluginConfig, options?: {
|
|
15
|
+
projectDirectory?: string;
|
|
16
|
+
}): AgentDefinition[];
|
|
15
17
|
/**
|
|
16
18
|
* Get agent configurations formatted for the OpenCode SDK.
|
|
17
19
|
* Converts agent definitions to SDK config format and applies classification metadata.
|
|
18
20
|
*
|
|
19
21
|
* @param config - Optional plugin configuration with agent overrides
|
|
22
|
+
* @param options - Optional options including projectDirectory
|
|
20
23
|
* @returns Record mapping agent names to their SDK configurations
|
|
21
24
|
*/
|
|
22
|
-
export declare function getAgentConfigs(config?: PluginConfig
|
|
25
|
+
export declare function getAgentConfigs(config?: PluginConfig, options?: {
|
|
26
|
+
projectDirectory?: string;
|
|
27
|
+
}): Record<string, SDKAgentConfig>;
|
|
23
28
|
/**
|
|
24
29
|
* Get the set of disabled agent names from config, applying protection rules.
|
|
25
30
|
*/
|
|
@@ -13,7 +13,7 @@ export interface AgentDefinition {
|
|
|
13
13
|
/**
|
|
14
14
|
* Resolve agent prompt from base/custom/append inputs.
|
|
15
15
|
* If customPrompt is provided, it replaces the base entirely.
|
|
16
|
-
*
|
|
16
|
+
* If customAppendPrompt is provided, it appends after whichever base won.
|
|
17
17
|
*/
|
|
18
18
|
export declare function resolvePrompt(base: string, customPrompt?: string, customAppendPrompt?: string): string;
|
|
19
19
|
/**
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
type AgentPermission = Record<string, 'allow' | 'ask' | 'deny' | Record<string, 'allow' | 'ask' | 'deny'>>;
|
|
2
|
+
/**
|
|
3
|
+
* Strict read-only tool permissions for advisory agents.
|
|
4
|
+
*
|
|
5
|
+
* Start with wildcard deny so newly-added tools are unavailable by default,
|
|
6
|
+
* then allow only inspection/search tools. Explicitly deny known mutating and
|
|
7
|
+
* delegation tools to make the read-only boundary obvious in generated config.
|
|
8
|
+
*/
|
|
9
|
+
export declare function createReadOnlyAgentPermission(): AgentPermission;
|
|
10
|
+
export {};
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* A custom skill bundled in this repository.
|
|
3
|
+
* Unlike npx-installed skills, these are copied from src/skills/ to the OpenCode skills directory
|
|
4
|
+
*/
|
|
5
|
+
export interface CustomSkill {
|
|
6
|
+
/** Skill name (folder name) */
|
|
7
|
+
name: string;
|
|
8
|
+
/** Human-readable description */
|
|
9
|
+
description: string;
|
|
10
|
+
/** List of agents that should auto-allow this skill */
|
|
11
|
+
allowedAgents: string[];
|
|
12
|
+
/** Source path in this repo (relative to project root) */
|
|
13
|
+
sourcePath: string;
|
|
14
|
+
}
|
|
15
|
+
/**
|
|
16
|
+
* Registry of custom skills bundled in this repository.
|
|
17
|
+
*/
|
|
18
|
+
export declare const CUSTOM_SKILLS: CustomSkill[];
|
|
@@ -1,21 +1,5 @@
|
|
|
1
|
-
|
|
2
|
-
|
|
3
|
-
* Unlike npx-installed skills, these are copied from src/skills/ to the OpenCode skills directory
|
|
4
|
-
*/
|
|
5
|
-
export interface CustomSkill {
|
|
6
|
-
/** Skill name (folder name) */
|
|
7
|
-
name: string;
|
|
8
|
-
/** Human-readable description */
|
|
9
|
-
description: string;
|
|
10
|
-
/** List of agents that should auto-allow this skill */
|
|
11
|
-
allowedAgents: string[];
|
|
12
|
-
/** Source path in this repo (relative to project root) */
|
|
13
|
-
sourcePath: string;
|
|
14
|
-
}
|
|
15
|
-
/**
|
|
16
|
-
* Registry of custom skills bundled in this repository.
|
|
17
|
-
*/
|
|
18
|
-
export declare const CUSTOM_SKILLS: CustomSkill[];
|
|
1
|
+
import { CUSTOM_SKILLS, type CustomSkill } from './custom-skills-registry';
|
|
2
|
+
export { CUSTOM_SKILLS, type CustomSkill };
|
|
19
3
|
/**
|
|
20
4
|
* Get the target directory for custom skills installation.
|
|
21
5
|
*/
|
|
@@ -23,7 +7,7 @@ export declare function getCustomSkillsDir(): string;
|
|
|
23
7
|
/**
|
|
24
8
|
* Install a custom skill by copying from src/skills/ to the OpenCode skills directory
|
|
25
9
|
* @param skill - The custom skill to install
|
|
26
|
-
* @param projectRoot - Root directory of oh-my-opencode-slim project
|
|
27
10
|
* @returns True if installation succeeded, false otherwise
|
|
11
|
+
* @deprecated Use syncBundledSkillsFromPackage instead.
|
|
28
12
|
*/
|
|
29
13
|
export declare function installCustomSkill(skill: CustomSkill): boolean;
|