immune

What problem does it solve?

Detects known errors and potential threats in generated content using an adaptive immune memory with HOT antibodies and COLD patterns, injects positive patterns before generation, and learns from results to improve future runs.

Core Features & Use Cases

• Hybrid memory system: Cheatsheet (positive patterns) injected before generation and Immune (negative patterns) scanned after generation.
• Hot/Cold memory management with per-scan caps to keep context lean and focused.
• Automated memory updates: corrections logging, threat detection, and new strategy learning for future scans.
• Multi-domain auto-detection and cross-domain scanning to handle complex content such as code, text, and data tasks.

Quick Start

Scan your content by providing it to Immune and let it automatically detect antibodies, log threats, and update its memory.