360 Secure SkillHub
360 安全技能中心

Zipsa

Zipsa 是位于 OpenClaw 与云端大模型之间的本地优先隐私网关：本地扫描并分类含 PII、健康记录、商业机密或凭据的请求，改写去除身份与机密细节后路由到云端模型，再把答复与本地上下文合成。支持通过会话 ID 续接多轮，并提供含策略矩阵、敏感词表、告警渠道与审计日志的管理控制台。

sulgik
sulgik
热度 54.28
最近更新 2026-05-11 17:34:09
分类：安全依赖：代码执行 / 第三方 API / 内部调用大模型应用场景：API 集成 / 数据处理 / 工作流

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add zipsa
Metadata
name
zipsa
description
Local-first privacy gateway for LLM applications. Use when the user needs to process sensitive, private, or proprietary information (PII, health records, internal business strategy, SSNs, credentials) while still utilizing cloud models (Claude, Gemini, GPT). Zipsa reformulates queries locally to remove identity and secrets before routing to the cloud, then synthesizes the answer back to the local context.

Zipsa

Zipsa is a local-first privacy gateway that sits between OpenClaw and cloud LLMs. It ensures that raw private context never leaves the local environment while still allowing the use of powerful frontier models for reasoning.

When to Use

Trigger this skill whenever the user provides or asks to process:

  • PII: Names, SSNs, DOBs, contact info.
  • Healthcare Data: Patient records, clinical notes, symptoms.
  • Internal Business Context: Pricing floors, negotiation terms, churn warnings, project codenames.
  • Secrets: API keys, credentials, internal system details.

Core Workflow

Zipsa operates as an OpenAI-compatible gateway (default: http://localhost:8000/v1).

1. Identify Sensitive Context

When the user's request contains sensitive data, explain that Zipsa will be used to protect their privacy.

2. Route Through Zipsa

Instead of calling the standard cloud providers directly, use the Zipsa endpoint. Zipsa will:

  • Scan locally for high-risk markers.
  • Classify the query (Private, Hybrid, or Knowledge).
  • Reformulate (rewrite) the query to abstract identity and proprietary details.
  • Route the safe version to the cloud.
  • Synthesize the final answer back with the local details.

Integration Examples

Using with OpenClaw

OpenClaw can be configured to use Zipsa as its backend:

  • OPENAI_BASE_URL=http://localhost:8000/v1
  • OPENAI_API_KEY=zipsa-key
  • OPENAI_MODEL=zipsa

Multi-turn Sessions

For conversation continuity, always pass a session_id.

{
  "model": "zipsa",
  "messages": [...],
  "extra_body": { "session_id": "unique-session-id" }
}

Admin Console

Zipsa includes an enterprise Admin Console for security teams and CISOs.

Key capabilities:

  • 실시간 대시보드: 차단/알람/마스킹 건수, 활성 에이전트 모니터링
  • 개인정보 유형 설정: 개인정보보호법 제2조 기준 (고유식별정보, 민감정보, 일반 개인정보, 금융정보, 인증정보)
  • 탐지·처리 정책 매트릭스: AI 서비스별 (ChatGPT/Claude/Gemini/Copilot) 처리 방식 설정
  • 조합 탐지 규칙: 단일 항목이 아닌 복합 식별자 조합 탐지
  • 민감정보 사전: 회사별 기밀 키워드 직접 등록 (CSV/XLSX 업로드 지원)
  • 알람 채널: Slack, 이메일, Teams, Webhook(SIEM/SOAR)
  • 인시던트 자동 대응: 임계값 기반 자동 차단 및 에스컬레이션
  • ISMS-P 컴플라이언스 매핑: 인증기준 2.0 (2024년 개정) 기준
  • 감사 로그: 3년 보존, 해시 체인 무결성 검증, SIEM 실시간 전송
  • 통계 리포트: 일/주/월간 보고서 자동 발송

For full Admin Console spec, see references/admin-dashboard.md.

Reference

For detailed configuration and advanced examples, see references/README.md.