4 条规则
# Zero-X security and analysis best practices
- Always consider running **zerox_full_scan** on generated, added, and modified first-party code, especially in security-sensitive areas (auth, crypto, payments, secrets handling, infrastructure-as-code).
- When running **zerox_full_scan**, provide the full relevant file contents and any important configuration files so Zero-X can produce accurate findings.
- When Zero-X reports issues, prefer using the remediation guidance returned by the tool to fix the code, and clearly explain how the fix addresses the underlying problem.
- After applying fixes based on Zero-X findings, run **zerox_full_scan** again on the affected files or project to verify that all reported issues have been resolved and no new issues were introduced.
- Prioritize remediation of **critical** and **high** severity findings first, then address medium and low severity issues as time permits.
- If the **Zero-X MCP server** is not installed or cannot be started successfully, inform the user and direct them to the official Zero-X MCP and Cursor plugin setup documentation.相关插件
zero-x-cursor↓ 4
Zero-X MCP 服务器将 Zero-X 的安全与分析能力直接带入 Cursor 中的 AI 编码工作流。no-secrets↓ 44
no-secrets:检测代码中可能的密钥泄露,建议环境变量/Secret Manager 修复,推送前警告。Acutis↓ 6
对 AI 生成代码进行零信任安全验证,详情见 https://acutis.dev。filestack↓ 6
Filestack 官方插件,可直接在 Cursor 中上传文件、构建转换并生成安全策略。contrastapi↓ 1
面向 AI agent 的安全情报 MCP 服务器,含 47 个工具、7 个资源与 3 个 prompt:CVE/EPSS/KEV(34 万+)、MITRE CWE/ATLAS/D3FEND、域名侦察、IOC 威胁情报、OSINT、代码安全与网络情报,免费且无需 API key。operant-mcp↓ 1
安全测试 MCP server,含 51 个工具,用于渗透测试、网络取证、内存分析与漏洞评估。ZPL Engine MCP↓ 0
面向金融、游戏、AI、加密的 AIN 偏差与稳定性引擎,提供 11 个类别共 68 个工具,进行数学中立性评分(0.1-99.9),用于组合平衡、掉落表公平性、模型偏差、大户集中度与合规评分等,免费版每月 5K token。Go↓ 4.5k
Go 在 Cursor 中的规则与最佳实践。