create-squirrel-opencode-harness 1.0.0 → 1.0.1
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.md +89 -29
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -14,13 +14,20 @@
|
|
|
14
14
|
- 🌐 Multi-language support (English/Chinese)
|
|
15
15
|
- 📁 Automatic `.opencode/` directory structure
|
|
16
16
|
|
|
17
|
+
## Prerequisites
|
|
18
|
+
|
|
19
|
+
Before installing, ensure you have:
|
|
20
|
+
|
|
21
|
+
- **Opencode CLI installed** - Install globally: `npm i -g opencode-ai`
|
|
22
|
+
- **AI Provider configured** - Set up your API keys in Opencode (OpenAI, Anthropic, Fireworks, etc.)
|
|
23
|
+
|
|
17
24
|
## Quick Start
|
|
18
25
|
|
|
19
|
-
No installation needed! Use `npm
|
|
26
|
+
No installation needed! Use `npm create` or `npx`:
|
|
20
27
|
|
|
21
28
|
```bash
|
|
22
|
-
# Using npm
|
|
23
|
-
npm
|
|
29
|
+
# Using npm create (recommended)
|
|
30
|
+
npm create squirrel-opencode-harness "fireworks-ai/accounts/fireworks/routers/kimi-k2p5-turbo"
|
|
24
31
|
|
|
25
32
|
# Using npx
|
|
26
33
|
npx create-squirrel-opencode-harness "fireworks-ai/accounts/fireworks/routers/kimi-k2p5-turbo"
|
|
@@ -68,22 +75,20 @@ echo "your-model-id" | create-squirrel-opencode-harness --stdin
|
|
|
68
75
|
create-squirrel-opencode-harness --interactive
|
|
69
76
|
```
|
|
70
77
|
|
|
71
|
-
### With npm
|
|
72
|
-
|
|
73
|
-
When using `npm init`, pass arguments after `--`:
|
|
78
|
+
### With npm create
|
|
74
79
|
|
|
75
80
|
```bash
|
|
76
|
-
# Basic usage with npm
|
|
77
|
-
npm
|
|
81
|
+
# Basic usage with npm create
|
|
82
|
+
npm create squirrel-opencode-harness "your-model-id"
|
|
78
83
|
|
|
79
84
|
# With language option
|
|
80
|
-
npm
|
|
85
|
+
npm create squirrel-opencode-harness "your-model-id" --lang zh
|
|
81
86
|
|
|
82
87
|
# Interactive mode
|
|
83
|
-
npm
|
|
88
|
+
npm create squirrel-opencode-harness --interactive --lang zh
|
|
84
89
|
|
|
85
90
|
# Using stdin (pipe model id)
|
|
86
|
-
echo "your-model-id" | npm
|
|
91
|
+
echo "your-model-id" | npm create squirrel-opencode-harness --stdin
|
|
87
92
|
```
|
|
88
93
|
|
|
89
94
|
### Language Selection
|
|
@@ -140,13 +145,38 @@ Options:
|
|
|
140
145
|
└── sprint-status-template.md
|
|
141
146
|
```
|
|
142
147
|
|
|
148
|
+
## Next Steps (After Installation)
|
|
149
|
+
|
|
150
|
+
Once the harness is scaffolded, follow these steps to start working:
|
|
151
|
+
|
|
152
|
+
### 1. Start Opencode
|
|
153
|
+
|
|
154
|
+
```bash
|
|
155
|
+
# In the same directory where you ran the scaffold
|
|
156
|
+
opencode
|
|
157
|
+
```
|
|
158
|
+
|
|
159
|
+
### 2. Switch to Harness Agent
|
|
160
|
+
|
|
161
|
+
- Press **Tab** to cycle through available agents
|
|
162
|
+
- Select **`harness`** agent - this is your main interface for AI collaboration
|
|
163
|
+
- ⚠️ **Important**: Always use the `harness` agent for communication. Do NOT use `planner`, `generator`, or `evaluator` directly - the harness will orchestrate these automatically.
|
|
164
|
+
|
|
165
|
+
### 3. Begin Your First Sprint
|
|
166
|
+
|
|
167
|
+
```
|
|
168
|
+
You: I want to build a todo list app
|
|
169
|
+
|
|
170
|
+
Harness: [Will create a sprint contract and guide you through the process]
|
|
171
|
+
```
|
|
172
|
+
|
|
143
173
|
## Examples
|
|
144
174
|
|
|
145
175
|
### Example 1: Quick Start
|
|
146
176
|
|
|
147
177
|
```bash
|
|
148
|
-
# Using npm
|
|
149
|
-
npm
|
|
178
|
+
# Using npm create (no install needed)
|
|
179
|
+
npm create squirrel-opencode-harness "openai/gpt-4"
|
|
150
180
|
|
|
151
181
|
# Or with npx
|
|
152
182
|
npx create-squirrel-opencode-harness "openai/gpt-4"
|
|
@@ -155,7 +185,7 @@ npx create-squirrel-opencode-harness "openai/gpt-4"
|
|
|
155
185
|
### Example 2: Interactive Mode in Chinese
|
|
156
186
|
|
|
157
187
|
```bash
|
|
158
|
-
npm
|
|
188
|
+
npm create squirrel-opencode-harness --interactive --lang zh
|
|
159
189
|
|
|
160
190
|
# Or with npx
|
|
161
191
|
npx create-squirrel-opencode-harness --interactive --lang zh
|
|
@@ -165,7 +195,7 @@ npx create-squirrel-opencode-harness --interactive --lang zh
|
|
|
165
195
|
|
|
166
196
|
```bash
|
|
167
197
|
export MODEL_ID="anthropic/claude-3-sonnet"
|
|
168
|
-
echo $MODEL_ID | npm
|
|
198
|
+
echo $MODEL_ID | npm create squirrel-opencode-harness --stdin
|
|
169
199
|
```
|
|
170
200
|
|
|
171
201
|
## Error Handling
|
|
@@ -204,13 +234,20 @@ MIT License - see [LICENSE](LICENSE) file for details.
|
|
|
204
234
|
- 🌐 多语言支持(英文/中文)
|
|
205
235
|
- 📁 自动创建 `.opencode/` 目录结构
|
|
206
236
|
|
|
237
|
+
## 前置要求
|
|
238
|
+
|
|
239
|
+
安装前,请确保您已具备:
|
|
240
|
+
|
|
241
|
+
- **Opencode CLI 已安装** - 全局安装:`npm i -g opencode-ai`
|
|
242
|
+
- **AI Provider 已配置** - 在 Opencode 中设置您的 API 密钥(OpenAI、Anthropic、Fireworks 等)
|
|
243
|
+
|
|
207
244
|
## 快速开始
|
|
208
245
|
|
|
209
|
-
无需安装!直接使用 `npm
|
|
246
|
+
无需安装!直接使用 `npm create` 或 `npx`:
|
|
210
247
|
|
|
211
248
|
```bash
|
|
212
|
-
# 使用 npm
|
|
213
|
-
npm
|
|
249
|
+
# 使用 npm create(推荐)
|
|
250
|
+
npm create squirrel-opencode-harness "fireworks-ai/accounts/fireworks/routers/kimi-k2p5-turbo"
|
|
214
251
|
|
|
215
252
|
# 使用 npx
|
|
216
253
|
npx create-squirrel-opencode-harness "fireworks-ai/accounts/fireworks/routers/kimi-k2p5-turbo"
|
|
@@ -258,22 +295,20 @@ echo "你的模型id" | create-squirrel-opencode-harness --stdin
|
|
|
258
295
|
create-squirrel-opencode-harness --interactive
|
|
259
296
|
```
|
|
260
297
|
|
|
261
|
-
### 使用 npm
|
|
262
|
-
|
|
263
|
-
使用 `npm init` 时,在 `--` 后传递参数:
|
|
298
|
+
### 使用 npm create
|
|
264
299
|
|
|
265
300
|
```bash
|
|
266
301
|
# 基础用法
|
|
267
|
-
npm
|
|
302
|
+
npm create squirrel-opencode-harness "your-model-id"
|
|
268
303
|
|
|
269
304
|
# 指定语言
|
|
270
|
-
npm
|
|
305
|
+
npm create squirrel-opencode-harness "your-model-id" --lang zh
|
|
271
306
|
|
|
272
307
|
# 交互模式
|
|
273
|
-
npm
|
|
308
|
+
npm create squirrel-opencode-harness --interactive --lang zh
|
|
274
309
|
|
|
275
310
|
# 使用标准输入(管道输入模型id)
|
|
276
|
-
echo "your-model-id" | npm
|
|
311
|
+
echo "your-model-id" | npm create squirrel-opencode-harness --stdin
|
|
277
312
|
```
|
|
278
313
|
|
|
279
314
|
### 语言选择
|
|
@@ -330,13 +365,38 @@ create-squirrel-opencode-harness "模型id" --lang zh
|
|
|
330
365
|
└── sprint-status-template.md
|
|
331
366
|
```
|
|
332
367
|
|
|
368
|
+
## 后续步骤(安装后)
|
|
369
|
+
|
|
370
|
+
脚手架搭建完成后,按以下步骤开始工作:
|
|
371
|
+
|
|
372
|
+
### 1. 启动 Opencode
|
|
373
|
+
|
|
374
|
+
```bash
|
|
375
|
+
# 在运行脚手架的同一目录
|
|
376
|
+
opencode
|
|
377
|
+
```
|
|
378
|
+
|
|
379
|
+
### 2. 切换到 Harness Agent
|
|
380
|
+
|
|
381
|
+
- 按 **Tab** 键循环切换可用 agent
|
|
382
|
+
- 选择 **`harness`** agent - 这是您与 AI 协作的主要界面
|
|
383
|
+
- ⚠️ **重要**:请始终使用 `harness` agent 进行交流。**不要**直接使用 `planner`、`generator` 或 `evaluator` - harness 会自动协调这些 agent。
|
|
384
|
+
|
|
385
|
+
### 3. 开始您的第一个 Sprint
|
|
386
|
+
|
|
387
|
+
```
|
|
388
|
+
您:我想构建一个待办事项应用
|
|
389
|
+
|
|
390
|
+
Harness: [将创建 sprint 合同并引导您完成整个过程]
|
|
391
|
+
```
|
|
392
|
+
|
|
333
393
|
## 使用示例
|
|
334
394
|
|
|
335
395
|
### 示例 1:快速开始
|
|
336
396
|
|
|
337
397
|
```bash
|
|
338
|
-
# 使用 npm
|
|
339
|
-
npm
|
|
398
|
+
# 使用 npm create(无需安装)
|
|
399
|
+
npm create squirrel-opencode-harness "openai/gpt-4"
|
|
340
400
|
|
|
341
401
|
# 或使用 npx
|
|
342
402
|
npx create-squirrel-opencode-harness "openai/gpt-4"
|
|
@@ -345,7 +405,7 @@ npx create-squirrel-opencode-harness "openai/gpt-4"
|
|
|
345
405
|
### 示例 2:中文交互模式
|
|
346
406
|
|
|
347
407
|
```bash
|
|
348
|
-
npm
|
|
408
|
+
npm create squirrel-opencode-harness --interactive --lang zh
|
|
349
409
|
|
|
350
410
|
# 或使用 npx
|
|
351
411
|
npx create-squirrel-opencode-harness --interactive --lang zh
|
|
@@ -355,7 +415,7 @@ npx create-squirrel-opencode-harness --interactive --lang zh
|
|
|
355
415
|
|
|
356
416
|
```bash
|
|
357
417
|
export MODEL_ID="anthropic/claude-3-sonnet"
|
|
358
|
-
echo $MODEL_ID | npm
|
|
418
|
+
echo $MODEL_ID | npm create squirrel-opencode-harness --stdin
|
|
359
419
|
```
|
|
360
420
|
|
|
361
421
|
## 错误处理
|