Safety Python Packages

nemoguardrails

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

cpex

A composable enforcement framework for AI agents and toolchains

skjold

Security audit Python project dependencies against security advisory databases.

safetext

Fast profanity word, curse word, swear word, bad word filtering tool for English, Spanish, Chinese, Turkish and more.

cato-cli-py

Portable sandbox for AI agents and untrusted commands. One config, any process, anywhere.

acgs-lite

Constitutional governance infrastructure for AI agents — the missing safety layer between your LLM and production

shadowaudit

Runtime governance for AI agents — deterministic fail-closed enforcement. Wraps any agent tool and blocks dangerous calls before execution. Zero LLM calls, zero cloud dependencies, works offline.

loom-agent

Python agent framework with AgentHarness: planner → generator ⇄ evaluator loops for long-running tasks—plus sessions, tool/safety policies, and Claude, GPT, or Gemini.

turboprivate-ai

Private AI / On-Premise AI Infrastructure

socratic-morality

Constitutional AI governance framework for ethical decision-making in agent systems

rigel-hpl

High-Level Property Specification Language

jmo-security

JMo Security Suite - Terminal-first security audit toolkit with many tools, multi-target scanning, & compliance

capfence

Runtime governance for AI agents — deterministic fail-closed enforcement. Wraps any agent tool and blocks dangerous calls before execution. Zero LLM calls, zero cloud dependencies, works offline.

agent-integrity-proto

Agent Integrity Protocol — real-time thinking block analysis for AI agent alignment

model-action-protocol

MAP (Model Action Protocol) — verifiable provenance, self-healing critique, and state rollback for autonomous AI agents

agentarmor

The full-stack safety layer for AI agents. Budget limits, prompt injection shields, PII filtering, output firewalls, and hooks — in 2 lines of code.

hpl-specs

High-Level Property Specification Language

phionyx-core

Deterministic AI runtime governance for LLM agents — treating model output as measurement, not authority. Python SDK with 46-block pipeline, signed audit trail, kill switch, ethics gates. PyPI: phionyx-core. AGPL-3.0.

ghostloop

The agent loop, embodied. Tool-using runtime + fail-closed safety pipeline + sim-first execution + statistically-rigorous bench harness + post-hoc analysis (counterfactual replay, causal attribution, LLM judge, property mining) for embodied AI. Sister project to GhostLM.

sh-guard

Semantic shell command safety classifier — AST-based risk scoring for AI coding agents

fmdtools

System resilience modelling, simulation, and assessment in Python

markdiffusion

MarkDiffusion: An Open-Source Toolkit for Generative Watermarking of Latent Diffusion Models

proofagent

pytest for AI agents — test safety, accuracy, tool use, and cost

simplesecurity

Combine multiple popular python security tools and generate reports or output into different formats