Add Sentinel AI to Red Teaming and Prompt Security

[MaxwellCalkin]

Summary

Adds Sentinel AI to the Red Teaming and Prompt Security section.

Sentinel AI is an open-source, real-time safety guardrails library for LLM applications:

• Prompt injection detection across 12 languages (English, Chinese, Japanese, Korean, Arabic, Hindi, etc.) with cross-lingual attack detection
• PII leak prevention — detects and redacts SSNs, credit cards, emails, phone numbers, and more
• Harmful content scanning — violence, self-harm, illegal activity, CSAM
• Obfuscation evasion detection — base64 encoding, zero-width character smuggling, Unicode homoglyphs, leetspeak
• Sub-millisecond latency (~0.05ms average scan time)
• Zero ML dependencies — pure pattern-based scanning with only regex as a dependency
• 530-case benchmark suite with 100% accuracy (0 false positives, 0 false negatives)
• Apache 2.0 licensed

It fits naturally alongside existing entries like LLM Guard, NeMo Guardrails, and Rebuff in the security tools table.

Checklist

• Entry follows existing table format (| **Name** | Description | [GitHub](url) |)
• Added to the appropriate section (Red Teaming and Prompt Security)
• Link is valid and points to a public repository