Repo · Safety & guardrails

Pi Warden

Flag risky actions and unverified completion.

Open github.com ↗

From the repository

Guardrails for Pi that steer instead of interrupt: enforces your project rules on every write, holds only hard-to-undo actions (3 per 1,000 calls, measured on real sessions), catches unverified "done" claims and stuck loops, and trims tool output. Built on pi-typesafe.

agent-guardrailsguardrailspi-extensionpi-packagetypesafe

How builders describe it

Guardrails for Pi built on pi-typesafe that steer the agent instead of interrupting you: Jev judges irreversible and off-task tool calls, detects stuck loops, checks unverified done claims, flags slop.

A Pi coding-agent extension that combines local checks with Jev judgments to hold selected actions, flag project-rule violations and steer stalled work.

i've been using something i built since yesterday morning up to now. it gives my coding agent a conscience, every action gets judged by jev in 250ms and it steers the agent accordingly so far i think jev is amazing to use together with LLMs 🤔 did anyone else built something similar im curious i…

The decision Jev makes

Allow, flag or escalate a message, tool call or action.

Where it fits

Screening messages, tool calls and outputs before they cause trouble. Jev gives a fast, calibrated verdict that can sit in front of every request or agent action. All 131 safety & guardrails projects →

Related projects