可信 — 风险评分 5/100
上次扫描:18 小时前 重新扫描
5 /100
inference-audit
Compare AI inference costs across providers and benchmark with real data
This is a legitimate cost-comparison skill that provides documentation and curl-based workflows for comparing AI inference pricing across providers, with no hidden functionality or security concerns.
技能名称inference-audit
分析耗时26.7s
引擎pi
可以安装
This skill is safe to use. Users should be aware they are interacting with a commercial service (GPU-Bridge) and that API calls include standard UTM tracking parameters.
资源类型声明权限推断权限状态证据
文件系统 NONE NONE No file operations found in documentation
网络访问 NONE READ ✓ 一致 Read-only GET requests to api.gpubridge.io for pricing data
命令执行 NONE NONE curl commands in docs are for user reference only, not executed by the skill
环境变量 NONE NONE No environment variable access detected
技能调用 NONE NONE No nested skill invocation
剪贴板 NONE NONE No clipboard access
浏览器 NONE NONE No browser automation
数据库 NONE NONE No database access
12 项发现
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=llm-4090
SKILL.md:48
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=embedding-l4
SKILL.md:49
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=whisper-l4
SKILL.md:50
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=image-4090
SKILL.md:51
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=tts-l4
SKILL.md:52
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=rerank
SKILL.md:53
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog
SKILL.md:56
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/account/register
SKILL.md:93
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/run
SKILL.md:98
🔗
中危 外部 URL 外部 URL
https://gpubridge.io
SKILL.md:133
🔗
中危 外部 URL 外部 URL
https://api.gpubridge.io/catalog/estimate?service=
tool.json:20
📧
提示 邮箱 邮箱地址
[email protected]
SKILL.md:95

目录结构

2 文件 · 6.8 KB · 167 行
Markdown 1f · 133L JSON 1f · 34L
├─ 📝 SKILL.md Markdown 133L · 5.7 KB
└─ 📋 tool.json JSON 34L · 1.2 KB

安全亮点

✓ No code execution, scripts, or binaries present - purely documentation-based skill
✓ Network access is read-only (GET requests for pricing data)
✓ No credential harvesting, key theft, or environment variable enumeration
✓ No obfuscation techniques (base64, eval, or anti-analysis patterns)
✓ No sensitive path access (~/.ssh, ~/.aws, .env files)
✓ No persistence mechanisms or startup hooks
✓ Transparent about commercial affiliation (GPU-Bridge marketing with UTM tracking)
✓ tool.json defines two safe, read-only tools with proper descriptions