Rune Journal

Field notes from building an agent security platform.

Research, threat reports, and engineering posts from the team. Written for the developers actually shipping AI agents, not for a buyer's committee.

FeaturedGuide·April 2026·16 min read

How to Secure Your AI Agent in Python: A Step-by-Step Guide

Secure a Python AI agent from scratch with input validation, output scanning, tool call policies, PII detection, and runtime monitoring. Working code for LangChain, OpenAI, Anthropic, and MCP.

By Declan Paul

Read article
Blog | Rune