Runtime integrity guard: detect and block Linux page cache tampering (Copy Fail, Dirty Pipe, Dirty Frag) via O_DIRECT + fanotify. Includes 7 host-side exploitation PoCs.
-
Updated
May 11, 2026 - Python
Runtime integrity guard: detect and block Linux page cache tampering (Copy Fail, Dirty Pipe, Dirty Frag) via O_DIRECT + fanotify. Includes 7 host-side exploitation PoCs.
🛡️ The CrowdStrike for AI Agents — Runtime protection, threat detection & security monitoring for LLM agents. Supports LangChain, CrewAI, AutoGen, OpenAI. Python • JS • Rust • Go • Ruby • .NET
Control plane for enterprise AI security: inline runtime guardrails, automated red teaming with auto-remediation, AI asset posture (AI-SPM), and compliance governance. Python control plane + Go inline agent.
Runtime guardrail for AI-generated Python code. Blocks rm -rf, curl | bash, and unapproved network calls before execution.
Enforces strict syscall whitelisting and egress IP cidr blocks for sandboxed autonomous code execution environments | GenPark: https://www.genpark.ai | LinkedIn: https://www.llinkedin.com/company/92955550
Security scanner and runtime protection for MCP (Model Context Protocol) servers
Deterministic in-process AST security runtime, secret zeroization, and Ed25519 Merkle receipts for Cursor, Claude Code, and autonomous AI agents (<35µs latency).
eBPF based Personal Runtime Protection - an open-source, and very basic client-server program designed to give your linux based computer the power of runtime protection using eBPF.
A deterministic runtime security SDK for LLM applications that prevents prompt injection, data leakage, and rogue agent behavior using high-performance, auditable rule-based guards instead of probabilistic AI inference.
The runtime firewall for AI agents. Protect any agent in 3 lines of code.
To associate your repository with the runtime-protection topic, visit your repo's landing page and select "manage topics."