Red Teaming Python Packages

dirsearch

Web path scanner

veldt-kya

KYA (Know Your Agents) — Open-source trust, governance, and evidentiary assurance infrastructure for autonomous systems. Built on KYP (Know Your Principal), a unified trust model for human users, AI agents, service accounts, and machine identities.

aicu-scanner

Black-box security scanner for LLM applications — prompt injection, safety bypass, credential leakage

agentic-radar

A security scanner for your LLM agentic workflows

crucible-security

pytest for AI agents - Autonomous red-teaming, behavioral monitoring & security testing for LLM agents

ziran

自然 ZIRAN is an open-source security testing framework for AI agents. It discovers dangerous tool chain compositions via knowledge graph analysis, detects execution-level side effects (not just text output), and runs multi-phase trust exploitation campaigns that model real attacker behaviour.

agentseal

Security toolkit for AI agents. Scan your machine for dangerous skills and MCP configs, monitor for supply chain attacks, test prompt injection resistance, and audit live MCP servers for tool poisoning.

pyadrecon-adws

An implementation of PyADRecon using ADWS instead of LDAP. Generates individual CSV files and a single XSLX + HTML report about your AD domain. Evades EDR detections through ADWS.

redharness

Standardized, reproducible LLM red-teaming benchmark across three surfaces (jailbreak · prompt-injection · data-leakage) under one methodology — judge-pinned provenance and a judge-sensitivity study. Offline-deterministic core; opt-in live model eval. Standardize the evaluation, not just the attack.

proofagent-harness

Open-source test harness for AI agents. Stress-test production agents with adversarial multi-turn scenarios in CI

helix-adapter

Portable constitutional adapter for any LLM. Forces epistemic markers ([FACT], [REASONED], etc.), real-time drift detection, and tamper-evident receipts. Model-agnostic wrapper.

pyadrecon

Python3 implementation of ADRecon with support for NTLM and Kerberos authentication querying LDAP. Generates individual CSV files and a single XSLX + HTML report about your AD domain.

xsafeclaw

Introducing XSafeClaw: The Open-Source Agent Safety Platform from Fudan University

aiverify-moonshot

Moonshot - A simple and modular tool to evaluate and red-team any LLM application.

rail-score-sdk

Official Python SDK for RAIL Score. LLM evaluation across 8 dimensions, guardrails, safe regeneration, agent tool-call eval, and compliance (GDPR, HIPAA, EU AI Act, India DPDP).

agent-release-gates

Release-readiness gates for AI agents: replay known incidents, apply policy-as-code, and produce ship/warn/block evidence before a prompt, model, or tool-policy change ships.

agentsec-bench

Security benchmark for LLM agents & their harnesses — offensive, defensive, secure-coding & prompt-injection-safety challenges on live sandboxed targets, with gradated scoring. Built on Inspect AI.

forcefield

ForceField Python SDK -- AI security in 3 lines of code. Prompt injection detection, PII redaction, security evals, tool governance. GitHub Action, pre-commit hook, Homebrew, VS Code extension.

guardrailprobe

Stop testing your model. Test your guardrail layer. guardrailprobe fires 78 OWASP LLM Top 10 attack probes across 11 backends — Lakera, NeMo, Presidio, Azure Content Safety, Bedrock Guardrails, LLM Guard, and five more — and outputs signed PDF and JSON comparison reports.

basilisk-ai

Basilisk — Open-source AI red teaming framework with genetic prompt evolution. Automated LLM security testing for GPT-4, Claude, Grok, Gemini. OWASP LLM Top 10 coverage. 32 attack modules.

rag-redteam

Red-team your RAG pipeline for prompt injection, source-document leakage, and cross-document smuggling, in CI. The gap between RAG eval (RAGAS/DeepEval) and LLM scanners (garak).

redlite

Opinionated tool for benchmarking Conversational Language Models

artkit

Automated prompt-based testing and evaluation of Gen AI applications

hermia

Vendor-agnostic security evals for LLM inference stacks. Detects behavioral divergence across models, backends, and hardware.