AI Security

Part 3: Resilient Tool Calling — Model Context Protocol (MCP) & Sandboxing

← Previous Chapter: Part 2: Hierarchical Memory | Series Hub | Next Chapter: Part 4: AgentOps & Observability → Answer-first: Standardizing agent tools on the Model Context Protocol (MCP) provides type-safe JSON-RPC contracts, token-budget enforcement, and secure capability boundaries. Code execution tools must run inside isolated WebAssembly (WASI 0.2) or micro-VM sandboxes.

Part 5: AI Code Security — OWASP LLM Top 10 & Supply-Chain Hardening

← Previous Chapter: Part 4: Multi-Agent Review Pipelines | Series Hub | Next Chapter: Part 6: Governance & Careers → Answer-first: Defending against AI code security risks requires automated package lockfile verification (blocking unregistered npm/PyPI packages) and strict secret masking pre-commit hooks to ensure private credentials never reach LLM context windows.

Part 6: Human-in-the-Loop (HITL) Gateways & Security Boundaries

← Previous Chapter: Part 5: Agent Evals | Series Hub Answer-first: For high-risk operations (financial fund transfers, database drop commands, production deployments), agents must pause execution state and request asynchronous human authorization through a durable workflow engine (Temporal / Dapr Workflows).

eBPF Zero-Trust Security for AI Agents: Tetragon 1.4

Tech Radar: eBPF Zero-Trust Security for AI Agents with Tetragon 1.4 Answer-First: Granting tool-execution permissions to AI Agents dramatically expands the attack surface for Remote Code Execution (RCE) via Indirect Prompt Injection. Cilium Tetragon 1.4 leverages eBPF probes inside the Linux kernel to intercept unauthorized system calls (execve, socket, openat), executing in-kernel SIGKILL enforcement in under 15 microseconds before malicious payloads can spawn reverse shells or exfiltrate credentials. 1. The Emerging Threat Vector: Autonomous Agent Prompt Injection RCE In modern agentic architectures, autonomous agents are granted tool execution permissions across the host environment: ...

NIST AI 600-1 & OWASP ASI01–ASI10: AI Gateways in Kubernetes

Tech Radar: NIST AI 600-1 & OWASP ASI01–ASI10 — Hardening Enterprise Agent Gateways in Kubernetes Answer-first: Deploying autonomous AI agent swarms into enterprise Kubernetes clusters demands a paradigm shift from Least Privilege to Least Agency. By unifying NIST AI 600-1 (the 12 GenAI Risk Categories across GOVERN/MAP/MEASURE/MANAGE) with the OWASP ASI Top 10 (2026 Agentic Security Standards), production architectures enforce a 4-tier defense: L7 Kubernetes Gateway API with CEL expressions for tool parameter sanitization, SPIFFE/SPIRE for ephemeral Non-Human Identity (NHI) mTLS attestation, and Cilium Tetragon eBPF for real-time Linux kernel syscall termination (SIGKILL < 15µs). ...