OpenRouter Launches Armor to Block Unauthorized AI Agent Tool Calls

OpenRouter launched Armor, a beta tool that uses decision models to block unauthorized AI agent tool calls before they interact with external systems. The product operates as an auto permissions mode compatible with any AI model or harness, utilizing a specialized classifier to replace traditional regex-based filtering. Armor is powered by Typesafe AI's Jev model, which OpenRouter says achieved higher accuracy and lower costs than other available decision models and the classifier used by Claude Code's Sonnet 5.