Llm Safety Python Packages

nemoguardrails

NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.

deepteam

DeepTeam is a framework to red team LLMs and LLM systems.

styxx

Cognitive observability for LLM agents. Cognometric instruments + self-healing reflex (F10) + MCP server. Pure-Python, MIT, no LLM required. 9-for-9 on K=1 phase transition. Every Mind Leaves Vitals (DOI 10.5281/zenodo.19777921).

agent-airlock

Open-source security firewall for AI agents — validates tool calls, strips ghost arguments, enforces type safety, PII masking, RBAC, cost tracking & sandbox isolation. Works with LangChain, OpenAI Agents SDK, PydanticAI & CrewAI.

uqlm

UQLM: Uncertainty Quantification for Language Models, is a Python package for UQ-based LLM hallucination detection

neurosym-ai

Neuro-symbolic guardrails for LLMs — injection detection, harm filters, output guards, streaming safety, and action-plan validation.

agent-audit

Static security scanner for LLM agents — prompt injection, MCP config auditing, taint analysis. 49 rules mapped to OWASP Agentic Top 10 (2026). Works with LangChain, CrewAI, AutoGen.

qwed-finance

Deterministic verification middleware for banking and financial AI. NPV, IRR, loan amortization, and interest calculations with QWED precision.

kitelogik

Governance control plane for autonomous AI agents — OPA/Rego policy enforcement at the agent action layer. Apache 2.0.

gateguard-ai

A fact-forcing hook gate for Claude Code. Makes the AI pause and investigate before editing.

csl-core

CSL-Core: Deterministic Safety Layer for Probabilistic AI Systems

qwed

The Deterministic Verification Protocol for AI - 11 verification engines for math, logic, code, SQL, facts, images, and more. Now with Agentic Security Guards.

lintlang

Static linter for AI agent configs, tool descriptions, and system prompts with zero-LLM CI gating

brix-protocol

Runtime Reliability Infrastructure for LLM Pipelines — enforce deterministic rules, measure the Balance Index, and audit every decision.

openbias

Open Source Reliability Harness: Make your agents follow rules. One line of code to‎ ‎enforce, trace, and improve. ‎ ‎

agi-pragma

AI Action Firewall — seven-stage Decision Intelligence Core for safe agentic AI

guardex

Guardex - AI Control Plane for autonomous agents (closed source)

blackwall-llm-shield-python

Security middleware for Python LLM apps and services. Blocks prompt injection, masks PII, inspects outputs, and gates agent tools.

fivedrisk

Per-action AI agent risk scoring and governance. Deterministic 5D scoring, HITL gating, FinOps, Agent Cost Management, Markov drift, audit log. Apache-2.0.

unworldly-recorder

The flight recorder for AI agents. Tamper-proof, ISO 42001 + HIPAA-compliant audit trails for everything AI agents do on your system. File changes + shell commands + PHI detection + agent identity.

jailbreakeval

[NDSS'25 Best Technical Poster] A collection of automated evaluators for assessing jailbreak attempts.

opensentinel

Reliability layer for AI agents - monitors workflow adherence and intervenes when agents deviate

medguard-llm

AI-powered medical guardrails API — PHI detection, drug safety, hallucination checks

ioa-core

IOA Core — the open-source governance-first kernel for AI orchestration. Clean public repo with OSS-only code, docs, and releases.