#guardrails

Validate and retry LLM outputs for ruby_llm. Describe the JSON response you expect, fall back to a stronger model when the cheaper one fails the rules, and gate CI on regressions — all as one contract object per step.

Open-source data protection framework for AI agent workflows in n8n. Detects, tokenizes, and reversibly anonymizes sensitive data before it reaches your models.

Production-grade Harness Engineering Framework for multi-developer AI coding agents — guardrails, state persistence, direction alignment, and self-correction across 6 IDEs. Zero dependencies.

Deterministic workflow control CLI for LLM (Coding) agents

AI Coding Gatekeeper — Zero-config, real-time guardrails for AI-generated code. 零配置AI编码守门员，实时拦截架构/安全/回归问题。

Two-layer enforcement (pre-commit hook + CI mirror) for small teams using AI agents. Catches debug leaks, file growth, secrets, and forbidden patterns before they merge.

The governance layer LLM proxies promised as a TypeScript library you can read. Every call runs through one governed pipeline: rate limit → spend caps (per-user + global circuit breaker) → cache → schema-validated provider failover → usage log → LLM-judge No sidecar proxy to deploy.

Public TypeScript SDK and integration examples for Punk, the adaptive runtime for production AI agents.

Guardrails for AI agents, with a receipt. Blocks destructive tool calls before they run and signs every decision into a hash-chained record anyone can verify with an open-source tool.