README.md

metadata

title: CodeIntegrity
emoji: 🛡️
colorFrom: blue
colorTo: gray
sdk: static
pinned: false

CodeIntegrity

CodeIntegrity secures endpoint AI agents and connected-tool execution. For supported agents, security teams can inspect configurations, skills, and session activity and apply configured policy before tool calls run.

Our research covers indirect prompt injection, MCP tool poisoning, and agent skill supply chain risk.

• Product and runtime controls
• Research and technical articles
• PromptGuard model and independent benchmark methods and results
• DERAIL prompt injection research and swarm poisoning research
• CodeIntegrity's $5M seed round led by SYN Ventures