-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathskill_test.yaml
More file actions
71 lines (60 loc) · 1.9 KB
/
skill_test.yaml
File metadata and controls
71 lines (60 loc) · 1.9 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
# AI Skill 测试框架配置文件
# 文档: python -m skill_test --help
cli:
command: claude # Windows 改为 claude.cmd
base_args:
- "--print"
- "--dangerously-skip-permissions"
timeout: 1000 # 单个任务超时(秒,coding 模式默认至少 1000 秒)
git:
author_name: "AI Skill Tester"
author_email: "skill-test@local"
base_branch: main # worktree 基于哪个分支创建
branch_prefix: skill-test
auto_push: false
cleanup_on_finish: true
retry:
max_retries: 2
base_delay: 5
max_delay: 60
retry_on_timeout: true
retry_on_failure: false
timeout_increment_on_timeout: 300 # 超时重试时额外增加的秒数(5 分钟)
max_workers: 3 # 并行执行上限
output_dir: results # 报告输出目录
report_formats:
- text
- json
# - markdown
# - html
tasks:
- id: task_001
name: "Python 快速排序"
prompt: |
实现一个快速排序算法,要求:
1. 使用 Python
2. 包含类型标注和文档字符串
3. 包含单元测试
timeout: 1000
- id: task_002
name: "API 错误处理"
prompt: |
重构以下代码的错误处理逻辑,使其更加健壮:
```python
def fetch_data(url):
response = requests.get(url)
return response.json()
```
skills:
- name: baseline
# 不设置 system_prompt = 基线对照
- name: write-expert
system_prompt: "你是专业的代码编写专家,擅长生成高质量、结构清晰的代码。"
- name: review-expert
system_prompt: "你是专业的代码审查专家,擅长发现代码问题、提供改进建议。"
# 从文件加载 Skill(适合大型 Skill 定义)
# - name: project-rules
# skill_file: ".claude/skills/dl-cloud-project-rules/SKILL.md"
# ref_files:
# - "references/architecture-rules.md"
# - "references/api-and-domain-rules.md"