devwithjev
— reading now— views
Submit a build

Dual-Brain Policy and Injection Agent

Leo Tavares built a proof-of-concept agent pairing Jev’s System 1 reflexes with an LLM’s System 2 deliberation. Jev resolved a multi-clause policy and intercepted an adversarial prompt injection.

View on X time250ms; 110ms
Leo Tavares@LeoTava8𝕏
Got access to Jev and wanted to build a proof-of-concept for amazing things it makes possible. I built a Dual-Brain agent pairing System 1 reflexes (Jev) with System 2 deliberation (LLM). In the 8x run below, you can watch both brains firing live: - Jev resolves a multi-clause policy in 250ms with zero LLM tokens. - Jev intercepts an adversarial prompt injection in 110ms, halting execution before
Sep 18, 2026X postsView on X
An 8x run shows both brains firing live. Jev handled the policy with zero LLM tokens and halted execution when it intercepted the injection.

Also filed under Agents & browsers

  • Evaluates Tool-Using Agents with Jev in CI

    The project evaluates a real tool-using AI agent using code checks, a calibrated Jev judge, and an LLM judge. It is wired into CI.

  • Watchdog for AI Coding Agents

    Jev Watch is a watchdog for AI coding agents that stops Codex or OpenCode when they loop, stall, or drift off task, then resumes the same session with a correction.

  • Custom Agent Harnesses with Jev

    Elvis uses Jev to build custom harnesses and reports success with guardrails, routing, and verifiers.

  • Fraud Investigation Agent

    The author is building a fraud investigation agent using Jev and TigerGraph. Its flow goes from an alert through graph evidence, a Bayesian update, a pattern, Jev’s extra lookup, policy, a verdict, and human approval.