“guardrails”

Rendered guardrail framework: keep a repo's hard rules in one-rule-per-file GRL sources behind a compact router, and render only the rules a task needs.

EMBEDDED — Pairs success metrics with guardrail metrics to check for trade-offs.

Safety guardrails: runtime overrides for AI-safety plugin blocks, and per-language security review for python, javascript/typescript, and go.

AI Agent Guardrails MCP server - security layer

Cloudflare Workers MCP server: ai-guardrails

Local prompt regression tracker with six fictional cases and evidence for human review.

Python framework with input and output guards that detect LLM risks, plus structured output generation

Open-source Python library for adding programmable guardrails to LLM-based conversational applications

Guards AI models, apps and agents against prompt injection, data leakage and harmful outputs.

AGENTS.md from opensandbox-group/OpenSandbox: 118 lines covering Repository Map, Routing, Working Principles, Guardrails and more.

AGENTS.md from graphile/crystal: 250 lines covering Repository Overview, Guardrails, Always Allowed, Preferred Style and more.

AGENTS.md from adobe/react-spectrum: 73 lines covering Repo layout, Toolchain guardrails, Visual verification with Storybook and docs, Contributing and more.

Python framework for agents that discover MCP tools by themselves, with optional memory, guardrails, cache and tracing.

Python SDK for multi-agent workflows with handoffs, guardrails, sessions and built-in tracing, usable with OpenAI and other LLMs.

TypeScript and JavaScript framework for multi-agent workflows with tools, handoffs, guardrails, sandboxes and voice agents.

Sparkle Guardrails

thesparkle

AI code security and guardrails for Cursor, Claude, Windsurf, VS Code & more - ON EVERY PROMPT

ThumbGate

igorganapolsky

Deterministic pre-action checks, feedback capture, and MCP guardrails for AI coding agents in VS Code-compatible IDEs.

Pulse Framework

pulse-framework

Guardrails, checkpoints, and escalation for AI-assisted development with Cursor Agent Mode.

Deterministic pre-trade risk layer for autonomous AI trading agents on Robinhood Chain.

HumanMirror is an independent agent control plane for autonomous systems: capability routing, pre-execution safety, agent action guardrails, fleet policy…

Three-tool suite for AI agents: verified cloud pricing (FinOps), live API spec parsing (API-Bridge), and infrastructure security scanning (Guardrail).

Control plane for models, data, guardrails and observability, plus Agent Builder v2 for deploying dockerized agent apps.

AWS service for agents that split requests into steps, call company APIs and query knowledge bases; closed to new customers

Goal-driven AI agent builder that lets you pick an LLM, use your data and templates, and orchestrate agents with bots and APIs.

AI-native proxy and data plane for agent apps: agent orchestration, LLM routing, guardrails and traces.

AI gateway on the same platform as your APIs, with multi-provider routing, per-team budgets and guardrails.

Enterprise AI gateway for 1000+ LLMs with governance, observability, guardrails and an MCP gateway.

Wallets, virtual cards and stablecoin infrastructure that let AI agents hold money and pay for services

Observe, enforce and build identity controls into AI agents, with runtime policy checks before agent actions run.

NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B.