invariant-evaluation

Here is 1 public repository matching this topic...

Ocherokee / semantic-intent-scanner

Semantic intent evaluation for agentic systems. Detects instruction-layer risks including authority laundering, consent bypass, concealed execution, and semantic attack patterns in SKILL.md and agent configuration files.

mcp cybersecurity semantic-analysis ai-security ai-alignment prompt-injection llm-security prompt-security agentic-ai semantic-intent semantic-execution skillmd invariant-evaluation

Updated Sep 5, 2026
Python

Add this topic to your repo

To associate your repository with the invariant-evaluation topic, visit your repo's landing page and select "manage topics."

Learn more