Add ShellWard - AI Agent Security Middleware

[jnMetaCode]

Adding ShellWard to the Tools section.

ShellWard is an AI agent security middleware with 8-layer defense:

• Prompt injection detection (32 rules, EN + ZH)
• DLP-style data flow tracking (read PII → outbound send = blocked)
• Dangerous command blocking (rm -rf, reverse shells)
• PII/API key detection
• Built-in CVE database (17 entries)

Works as SDK for any agent framework or as OpenClaw plugin. Zero deps. Apache-2.0.

GitHub: https://github.com/jnMetaCode/shellward