Zipsa

Local-first privacy gateway for LLM applications. Use when the user needs to process sensitive, private, or proprietary information (PII, health records, int...

This skill is a configuration/routing helper that expects you to run a separate local Zipsa server. Before installing or enabling it: 1) Ensure you actually host a trusted Zipsa instance on localhost (the skill alone does not provide scanning/masking logic). 2) Do not set OPENAI_BASE_URL to a remote or untrusted URL — the gateway must be local/trusted to protect sensitive data. 3) Review and lock down admin-console integrations (Slack, webhooks, SIEM) because misconfigured alert/webhook destinations could leak sensitive events. 4) Test with non-sensitive sample data first to confirm behavior (blocking/masking) and inspect where audit logs are sent and stored. 5) If you need the gateway to contact cloud providers, provision provider credentials in a controlled location and review that those credentials are used only by your trusted Zipsa service.

Type: OpenClaw Skill Name: zipsa Version: 0.4.1 The 'zipsa' skill is a privacy-preserving gateway designed to intercept and sanitize sensitive data (PII, credentials, health records) locally before routing requests to cloud LLM providers. The provided Python script (scripts/relay_zipsa.py) is a straightforward relay to a local endpoint (localhost:8000), and the extensive documentation (SKILL.md, references/admin-dashboard.md) describes legitimate security features, including ISMS-P compliance mapping and local data masking, with no evidence of malicious intent, data exfiltration, or unauthorized execution.