메뉴
AI与MCP安全评估 — Prompt注入、工具滥用、MCP信任边界、Agent权限逃逸、数据泄露、模型风险、GAARM风险矩阵
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Domain routing and boundary guidance for authorized Active Directory red-team security testing, including Kerberos attacks, domain privilege escalation, lateral movement, and GPO abuse. Use when a task belongs to the AD testing domain and needs scope, evidence, pivot, or exit criteria.
Domain routing and boundary guidance for authorized API security testing, including BOLA/IDOR, authentication bypass, mass assignment, missing rate limits, and GraphQL issues. Use when a task belongs to the API testing domain and needs scope, evidence, pivot, or exit criteria.
Domain routing and boundary guidance for authorized authentication, authorization, and session security testing, including password policy, JWT/token, OAuth, and MFA bypass issues. Use when a task belongs to the auth testing domain and needs scope, evidence, pivot, or exit criteria.
Domain routing and boundary guidance for authorized web cache poisoning testing, including unkeyed headers, unkeyed parameters, cache deception, and CDN-specific behavior. Use when a task belongs to the cache poisoning domain and needs scope, evidence, pivot, or exit criteria.
Domain routing and boundary guidance for authorized clickjacking testing, including missing X-Frame-Options, CSP frame-ancestors bypasses, and drag-and-drop hijacking. Use when a task belongs to the clickjacking domain and needs scope, evidence, pivot, or exit criteria.
Domain routing and boundary guidance for authorized cloud security testing, including IAM misconfiguration, exposed storage, metadata services, and serverless injection. Use when a task belongs to the cloud testing domain and needs scope, evidence, pivot, or exit criteria.
| name | ai-mcp-security |
| description | AI与MCP安全评估 — Prompt注入、工具滥用、MCP信任边界、Agent权限逃逸、数据泄露、模型风险、GAARM风险矩阵 |
| routing | {"target_types":["ai_agent","mcp"],"task_types":["pentest","audit"],"technologies":["llm","rag","memory","plugin"],"vulnerability_classes":["prompt_injection","tool_abuse","info_disclosure"]} |
当目标包含 LLM、Agent、MCP 工具、Skills、RAG、Memory、Plugin 或模型服务组件时使用本 Skill。
前置条件：如果 AI 表面只是展示层，真正的阻塞仍是客户端签名或加密协议，先回到 client-reverse Skill。
| 风险类型 | 首选参考 |
|---|---|
| Prompt 注入 / 间接注入 / CoT 干扰 | references/ai-app-security.md |
| 工具滥用 / MCP 投毒 / Skills 供应链 | references/04-ai-and-mcp-security-integrated.md MCP 章节 |
| 权限逃逸 / 角色越界 / 凭据滥用 | references/ai-identity-security.md |
| 数据泄露 / Prompt 泄漏 / 模型逆推 | references/ai-data-security.md |
| 容器逃逸 / CI-CD / 沙箱失败 | references/ai-baseline-security.md |
| 模型风险 / 对抗样本 / 后门 | references/ai-model-security.md |
| 影响分类与覆盖评估 | references/gaarm-risk-matrix.md |
references/04-ai-and-mcp-security-integrated.md — AI 与 MCP 安全整合参考references/ai-app-security.md — AI 应用安全references/ai-identity-security.md — AI 身份安全references/ai-data-security.md — AI 数据安全references/ai-baseline-security.md — AI 基线安全references/ai-model-security.md — AI 模型安全references/gaarm-risk-matrix.md — GAARM 风险矩阵