opendcai/bench
.claude/skills/bench/SKILL.md
测试 API 模型性能 - TTFB 和吐字速度
$ install --global
skillsauthnpx skillsauth add opendcai/leonai benchInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 10, 2026, 9:23 AM5.8s3 files scanned
SKILL.md
- name:
- bench
- description:
- 测试 API 模型性能 - TTFB 和吐字速度
快速测试第三方 API 的模型性能。两个脚本，按 API 格式选择。
用法
用户提供：API Key + Base URL
默认 → OpenAI 格式（/v1/chat/completions）：
cd .claude/skills/bench && uv run python test_models.py "<API_KEY>" "<BASE_URL>"
用户说"Anthropic 格式"→ Anthropic 原生（/v1/messages）：
cd .claude/skills/bench && uv run python test_anthropic.py "<API_KEY>" "<BASE_URL>"
判断规则
- 默认走
test_models.py（OpenAI 格式） - 仅当用户明确说"Anthropic 格式/原生 API"时，走
test_anthropic.py - 不要自动猜测，由用户指定
测试指标
- TTFB: 首次响应延迟（10秒超时）
- Tok/s: 吐字速度
- 总超时：30秒
OpenAI 脚本特性（test_models.py）
- 自动从
/models端点获取可用模型并智能筛选 - 筛选最新主流文本模型（Claude 4.x, GPT 5.x, Gemini 3.x, Qwen 3.x, GLM 4.7+, Kimi k2.5+）
- 排除 DeepSeek、多模态模型、过时版本
- Base URL 必须带
/v1后缀
Anthropic 脚本特性（test_anthropic.py）
- 使用
x-api-key+anthropic-version认证 - 默认测试：opus-4-6, sonnet-4-6, sonnet-4-5, haiku-4-5
- Base URL 带不带
/v1都行（自动处理）
常见问题
- 依赖：必须用
uv run，不能直接python - OpenAI Base URL：必须带
/v1，否则/models返回 HTML - Anthropic
/models被拦截：正常，Cloudflare 保护，脚本用默认模型列表
Related Skills
opendcai/guiding-users
tools
VerifiedTrustedCommunity
Implements onboarding and help systems including product tours, interactive tutorials, tooltips, checklists, help panels, and progressive disclosure patterns. Use when building first-time experiences, feature discovery, guided walkthroughs, contextual help, setup flows, or user activation features. Provides timing strategies, accessibility patterns (keyboard, screen readers, reduced motion), and metrics for measuring onboarding success.
101SKILL.mdUpdated Apr 10, 2026
opendcai/frontend-design
development
VerifiedTrustedCommunity
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
101SKILL.mdUpdated Apr 10, 2026
opendcai/react
development
VerifiedTrustedCommunity
Builds token-driven React components with TypeScript and modern patterns. Use when creating React component libraries, integrating CSS custom properties, or building Next.js design system components with forwardRef and composition.
101SKILL.mdUpdated Apr 10, 2026
opendcai/event-store-design
development
VerifiedTrustedCommunity
Design and implement event stores for event-sourced systems. Use when building event sourcing infrastructure, choosing event store technologies, or implementing event persistence patterns.
101SKILL.mdUpdated Apr 10, 2026