Add PrivacyScrubber (client-side zero-trust PII & secret redaction guard) to AI section

[moxno]

Adding one tool to the AI section, following the repository contribution rules.

Tool: PrivacyScrubber (moxno/privacyscrubber-mcp)

What it is: A client-side Zero-Trust Data Sanitization (ZTDS) engine, Model Context Protocol (MCP) server, and CLI tool that redacts customer PII, enterprise credentials, and sensitive tokens locally in memory before prompt dispatch to LLMs and agentic workflows (Cursor, Claude Code, etc.), featuring zero network egress and deterministic reverse-unmasking.

Why: The AI section currently covers multi-cluster compliance observability (KubeStellar Console) and sandboxed research CLI (Cynative), but lacks an inline data-sanitization / DLP guard preventing prompt leakage of secrets and regulated data into AI pipelines. This covers a distinct functional category rather than duplicating existing tools.

Topic: AI (agentic security & zero-trust data sanitization).

Meta:

• License: Apache-2.0 (open source)
• Active: released on npm as @privacyscrubber/mcp-server (v2.2.1) and @privacyscrubber/sdk (v2.2.0)
• Stars: 2 (reference implementation open-sourced alongside npm releases)
• Security profile: 100% local RAM execution, 0 bytes network egress

One row added to the AI table, matching the existing Name | URL | Description | Meta format. No other lines touched.