Skip to main content

Correctover Conformance Standard - AI/MCP Agent Security Audit Engine

Project description

Correctover Conformance Standard (CCS)

AI/MCP Agent 安全审计引擎 — 从静态代码扫描到可验证漏洞报告的完整工具链。

核心能力

CCS 专为 AI Agent / MCP (Model Context Protocol) 生态设计,提供六层递进式安全分析:

模块 功能 解决的问题
Module 1 输入可控性分类 参数是管理员配置还是外部可控?
Module 2 数据流追踪 输入能否到达危险函数?
Module 3 隐式失败检测 攻击链会不会被中间件/框架偷偷拦截?
Module 4 MCP专项规则 MCP协议层特有的安全风险
Module 5 置信度评分 多维交叉验证,过滤误报
Module 6 报告生成 自动生成 MSRC / Bugcrowd / GitHub Advisory 格式报告

快速开始

pip install correctover-ccs

# 对目标项目执行完整审计
ccs /path/to/target/project --output ./audit_results/

流水线架构

源代码 → [M1: 输入分类] → [M2: 数据流追踪] → [M3: 隐式失败检测] → [M4: MCP规则]
                                                                          ↓
                                                                   [M5: 置信度评分]
                                                                          ↓
                                                                   [M6: 报告生成]
                                                                          ↓
                                                                    漏洞报告 (Markdown)

MCP专项规则覆盖

  • MCP-001: MCP HTTP URL零校验 → SSRF (CWE-918)
  • MCP-002: MCP Stdio命令零校验 → 命令注入 (CWE-78)
  • MCP-003: 工具注册无权限检查 (CWE-862)
  • MCP-004: Agent间数据隔离不足 (CWE-200)
  • MCP-005: 云元数据地址直接访问 (CWE-918)
  • MCP-006: 技能/插件下载无URL验证 → SSRF (CWE-918)
  • MCP-007: Prompt注入防护缺失 (CWE-74)
  • MCP-008: 工具调用参数未验证 (CWE-20)

置信度评分体系

总分100分,五维度加权:

  • 输入可控性 (30分): 外部可控=30, 认证用户=18, 管理员=6, 环境配置=3
  • 数据流可达性 (25分): 无阻断=25, 有阻断=10, 不可达=0
  • 隐式失败 (20分): 无失败=20, 1个失败点=10, 多个=4
  • MCP规则匹配 (15分): 按严重性等比映射
  • 影响严重性 (10分): 按CVSS等级映射

置信等级:HIGH(≥80) → 立即验证 | MEDIUM(50-79) → 安排验证 | LOW(<50) → 跳过

作为库使用

from ccs import CCSEngine

engine = CCSEngine("/path/to/target")
summary = engine.run_full_audit()
files = engine.export_results("./output/")

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

correctover_ccs-1.1.0.tar.gz (24.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

correctover_ccs-1.1.0-py3-none-any.whl (30.2 kB view details)

Uploaded Python 3

File details

Details for the file correctover_ccs-1.1.0.tar.gz.

File metadata

  • Download URL: correctover_ccs-1.1.0.tar.gz
  • Upload date:
  • Size: 24.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.12

File hashes

Hashes for correctover_ccs-1.1.0.tar.gz
Algorithm Hash digest
SHA256 b0ee2332707dcebac31d161699bb6f150641982a7df8a1b22d90580e75cd157b
MD5 5e4c2b05bc52394e3edd0cc89d3cccb3
BLAKE2b-256 ac78bba56bb72a4934e483007a11e9796b969577f5b6d47346ea9ba5072e1f5e

See more details on using hashes here.

File details

Details for the file correctover_ccs-1.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for correctover_ccs-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 4c6d5b7479b40f6fc23c0d740ee19f28a134c123f6aef345f3ebf24bcdbf64f9
MD5 9290886974149ecb205ab7074b05aa2e
BLAKE2b-256 3885222f4840a492a26d3f8068537a11d945af46d6566a2a6fff008cc511e761

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page