经 AI Skill Hub 精选评估,AI安全工具 获评「推荐使用」。这款MCP工具在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 7.5 分,适合有一定技术背景的用户使用。
AI安全工具 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
AI安全工具 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
# 方式一:通过 Claude Code CLI 一键安装
claude skill install https://github.com/quantifylabs/aegis-memory
# 方式二:手动配置 claude_desktop_config.json
{
"mcpServers": {
"ai----": {
"command": "npx",
"args": ["-y", "aegis-memory"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
# 安装后在 Claude 对话中直接使用 # 示例: 用户: 请帮我用 AI安全工具 执行以下任务... Claude: [自动调用 AI安全工具 MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入:"列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"ai____": {
"command": "npx",
"args": ["-y", "aegis-memory"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
<p align="center"> <img src=".github/banner.svg" alt="Aegis Memory" width="400"/> </p>
<p align="center"> <strong>Your agent's context is your attack surface. Act accordingly.</strong> </p>
<p align="center"> Secure context engineering for production AI agents.<br/> Content security. Integrity verification. Trust hierarchy. Context that improves itself. </p>
<p align="center"> <a href="https://github.com/quantifylabs/aegis-memory/actions/workflows/ci.yml"><img src="https://github.com/quantifylabs/aegis-memory/actions/workflows/ci.yml/badge.svg" alt="CI"></a> <a href="https://pypi.org/project/aegis-memory/"><img src="https://img.shields.io/pypi/v/aegis-memory?label=PyPI&color=blue" alt="PyPI version"></a> <a href="https://pypi.org/project/aegis-memory/"><img src="https://img.shields.io/pypi/dm/aegis-memory?label=downloads&color=brightgreen" alt="PyPI downloads"></a> <a href="https://pypi.org/project/aegis-memory/"><img src="https://img.shields.io/pypi/pyversions/aegis-memory.svg" alt="Python versions"></a> <a href="https://opensource.org/licenses/Apache-2.0"><img src="https://img.shields.io/badge/License-Apache_2.0-blue.svg" alt="License"></a> <a href="https://deepwiki.com/quantifylabs/aegis-memory"><img src="https://img.shields.io/badge/DeepWiki-quantifylabs%2Faegis--memory-blue.svg" alt="DeepWiki"></a> <a href="https://scorecard.dev/viewer/?uri=github.com/quantifylabs/aegis-memory"><img src="https://api.scorecard.dev/projects/github.com/quantifylabs/aegis-memory/badge" alt="OpenSSF Scorecard"></a> <a href="https://www.bestpractices.dev/projects/13029"><img src="https://www.bestpractices.dev/projects/13029/badge" alt="OpenSSF Best Practices"></a> <a href="https://github.com/quantifylabs/aegis-memory"><img src="https://img.shields.io/github/stars/quantifylabs/aegis-memory?style=social" alt="GitHub stars"></a> </p>
<p align="center"> <a href="https://www.aegismemory.com/">Website</a> • <a href="https://docs.aegismemory.com/introduction/overview">Docs</a> • <a href="https://www.aegismemory.com/blog/">Blog</a> • <a href="https://docs.aegismemory.com/quickstart/installation">Quickstart</a> • <a href="https://docs.aegismemory.com/guides/security">Security Guide</a> </p>
---
pip install aegis-memory
docker compose up -d
| Variable | Default | Description |
|---|---|---|
DATABASE_URL | postgresql+asyncpg://... | PostgreSQL connection |
OPENAI_API_KEY | — | For embeddings |
AEGIS_API_KEY | dev-key | API authentication |
CONTENT_POLICY_INJECTION | flag | reject / redact / flag / allow |
CONTENT_POLICY_SECRETS | reject | reject / redact / flag / allow |
ENABLE_LLM_INJECTION_CLASSIFIER | false | Enable Stage 4 LLM classifier |
INJECTION_CLASSIFIER_MODEL | gpt-4o-mini | Model for injection classification |
| Feature | ACE-Inspired | Aegis ACE-Engineered |
|---|---|---|
| Voting | Manual vote endpoints | Auto-voting tied to run outcomes |
| Reflection | Manual reflection creation | Auto-reflection on failure with error context |
| Curation | Not implemented | Full curation cycle with promote/flag/consolidate |
| Run tracking | Not tracked | First-class ace_runs table linking memories to outcomes |
| Agent-specific playbook | Generic query | Filtered by agent_id + task_type |
Memory-depth primitives (hybrid retrieval, contradiction handling, consolidation) are now table stakes — mem0, Zep, Letta, and Aegis all ship variants in 2026.[^memory-depth-sources] The differences are in how, not whether.
| Capability | mem0 | Graphiti / Zep | Letta | Aegis Memory |
|---|---|---|---|---|
| **Primary focus** | Assistant personalization | Graph-based episodic memory | Stateful agents | Secure context engineering |
| **Open source** | Yes | Yes | Yes | Yes |
| **Self-host posture** | Available | Available | Available | Self-host-first |
| **Content security pipeline** | — | — | — | 4-stage (validation, PII, injection, LLM) |
| **Memory integrity** | — | — | — | HMAC-SHA256 |
| **Trust hierarchy** | — | — | — | 4-tier OWASP model |
| **Multi-agent ACL/scopes** | — | — | — | Yes |
| **Cross-agent query** | — | — | — | Yes |
| **Handoff baton** | — | — | — | Yes |
| **ACE loop** | — | — | — | Yes |
| **Typed memory model** | — | — | — | Yes |
| **Temporal decay** | — | Partial | — | Yes |
| **Hybrid retrieval (dense + sparse + RRF)** | Semantic + BM25 + entity | Semantic + keyword + graph | Yes (RRF) | Yes (pgvector + tsvector + RRF) |
| **Contradiction detection** | Mem0g (graph variant, LLM) | LLM + temporal invalidation | — | Typed contradicts edge, cheap + optional LLM, **explicit resolution workflow** |
| **Semantic consolidation** | LLM-merge + DELETE losers | Temporal supersession | — | LLM/heuristic merge + **audit-preserving** (is_deprecated=True + consolidated_into) |
| **Unified context hub (prompts + memory + skills + subagents)** | — | — | — | Yes |
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ Apache 2.0 — 宽松开源协议,可商用,需保留版权声明和 NOTICE 文件,含专利授权条款。
AI Skill Hub 点评:AI安全工具 的核心功能完整,质量良好。对于Claude Desktop / Claude Code 用户来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。
| 原始名称 | aegis-memory |
| 原始描述 | 开源MCP工具:Secure context engineering for AI agents. Content security · integrity verificat。⭐22 · Python |
| Topics | aiagent-securitymcp |
| GitHub | https://github.com/quantifylabs/aegis-memory |
| License | Apache-2.0 |
| 语言 | Python |
收录时间:2026-05-31 · 更新时间:2026-06-02 · License:Apache-2.0 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端