Tools listed by the source
Compendium tool directory, snapshot 2026-10-22. Listings and descriptions are attributed source claims; no installation, evaluation, current availability or endorsement is implied. Some entries repeat across categories. Review the actual tool, license, data handling and assessment authorization before use.
| Category | Tool | Source description |
|---|---|---|
| Model Scanners | HiddenLayer Model Scanner | A tool that scans AI models to detect embedded malicious code, vulnerabilities, and integrity issues, ensuring secure deployment. |
| Model Scanners | Fickling | An open-source utility for analyzing and modifying Python pickle files, commonly used for serializing machine learning models. |
| Model Scanners | Protect AI Guardian | An enterprise-level tool that scans third-party and proprietary models for security threats before deployment, enforcing model security policies. |
| Model Scanners | AppSOC’s AI Security Testing solution | AppSOC’s AI Security Testing solution helps in proactively identifying, and assessing the risks from LLM models by automating model scanning, simulating adversarial attacks, and validating trust in connected systems, ensuring the models and ecosystems are safe, compliant, and deployment-ready |
| Model Scanners | SplxAI Probe | An enterprise platform for evaluating models for security, safety, trust, and alignment using a continuously updated attack database, custom attacks, and the latest multi-turn attack strategies. |
| Model Validation Tools | Robust Intelligence Continuous Validation | A platform offering continuous validation of AI models to detect and mitigate vulnerabilities, ensuring robust and secure AI deployments. |
| Model Validation Tools | Protect AI Recon | A product that automatically validates LLM Model performance across common industry framework requirements (OWASP, MITRE/ATLAS). |
| Model Validation Tools | Vigil LLM security scanner | A tool designed to scan large language models (LLMs) for security vulnerabilities, ensuring safe deployment and usage. |
| Model Validation Tools | Garak Automated Scanning | An automated system that scans AI models for potential security threats, focusing on detecting malicious code and vulnerabilities. |
| Model Validation Tools | HiddenLayer AIDR | A solution that monitors AI models in real time to detect and respond to adversarial attacks, safeguarding AI assets. |
| Model Validation Tools | Citadel Lens | A security tool that provides visibility into AI models, detecting vulnerabilities and ensuring compliance with security standards. |
| Model Validation Tools | AppSOC’s AI Security Testing solution | AppSOC’s AI Security Testing solution helps in proactively identifying, and assessing the risks from LLM models by automating model scanning, simulating adversarial attacks, and validating trust in connected systems, ensuring the models and ecosystems are safe, compliant, and deployment-ready |
| Model Validation Tools | SplxAI Probe | An end-to-end enterprise security suite for GenAI applications powered by the latest adversarial techniques, LLM-as-judge, AI-based remediation, log analysis, and compliance mapping to popular AI security frameworks. |
| AI Agents | Arhasi R.A.P.I.D | A platform offering rapid assessment and protection of AI deployments, focusing on identifying and mitigating security risks. |
| AI Agents | SplxAI Agentic Radar | Open-source tool to analyze agentic systems for security and operational insights. It helps developers, researchers, and security professionals understand how agentic systems function and identify potential vulnerabilities. |
| Guardrails for LLMs | NeMo Guardrails | A toolkit for adding programmable guardrails to AI models, ensuring they operate within defined safety and ethical boundaries. |
| Guardrails for LLMs | Guradrails AI | A framework that integrates safety protocols into AI models, preventing them from generating harmful or biased outputs. |
| Guardrails for LLMs | Lakera Guard | A security solution that monitors AI models for adversarial attacks and vulnerabilities, providing real-time protection. |
| Guardrails for LLMs | Robust Intelligence AI Firewall | A protective layer that shields AI models from adversarial inputs and attacks. |
| Guardrails for LLMs | Protect AI Layer | Layer provides LLM runtime security including observability, monitoring, blocking for AI Applications. The enterprise grade offering brought to you by the same team that built the industry leading open source solution LLM Guard. |
| Guardrails for LLMs | Arthur Shield | A monitoring solution that tracks AI model performance and security, detecting anomalies and potential threats in real time. |
| Guardrails for LLMs | Amazon Guardrails | A set of safety protocols integrated into Amazon's AI services to ensure models operate within ethical and secure boundaries. |
| Guardrails for LLMs | Meta Llama Guard | Meta implemented security measures to protect their Llama models from vulnerabilities and adversarial attacks. |
| Guardrails for LLMs | Arhasi R.A.P.I.D | A platform offering rapid assessment and protection of AI deployments, focusing on identifying and mitigating security risks. |
| Guardrails for LLMs | SplxAI Rail | Low-latency prompt and response monitoring and flagging for configurable forbidden topics related to security and safety. |
| DASF Validation and Assessment Products and Services | Safe Security | SAFE One makes cybersecurity an accelerator to the business by delivering the industry's only data-driven, unified platform for managing all your first-party and third-party cyber risks. |
| DASF Validation and Assessment Products and Services | Obsidian | Obsidian Security combines application posture with identity and data security, safeguarding SaaS. |
| DASF Validation and Assessment Products and Services | EQTY Labs | EQTY Lab builds advanced governance solutions to evolve trust in AI. |
| DASF Validation and Assessment Products and Services | AppSOC | Makes Databricks the most secure AI platform with real-time visibility, guardrails, and protection. |
| Public AI Red Teaming Tools | Garak | An automated scanning tool that analyzes AI models for potential security threats, focusing on detecting malicious code and vulnerabilities. |
| Public AI Red Teaming Tools | BlackIce | An open-source, containerized toolkit that bundles 14 widely used AI security tools into a single, reproducible environment. In this post, we highlight the motivation behind BlackIce, outline its core capabilities, and share resources to help you get started. |
| Public AI Red Teaming Tools | Protect AI Recon | A product with a full suite of Red Teaming options for AI applications, including a library of common attacks, human augmented attacks, and LLM generated scans; complete with mapping to common industry frameworks like OWASP and MITRE/ATLAS. |
| Public AI Red Teaming Tools | PyRIT | A Python-based tool for testing the robustness of AI models against adversarial attacks, ensuring model resilience. |
| Public AI Red Teaming Tools | Adversarial Robustness Toolbox (ART) | An open-source library that provides tools to assess and improve the robustness of machine learning models against adversarial threats. |
| Public AI Red Teaming Tools | Counterfeit | A tool designed to test AI models for vulnerabilities by simulating adversarial attacks, helping developers enhance model security. |
| Public AI Red Teaming Tools | ToolBench | A suite of tools for evaluating and improving the security and robustness of AI models, focusing on detecting vulnerabilities. |
| Public AI Red Teaming Tools | Giskard-AI llm scan | A tool that scans large language models for security vulnerabilities, ensuring safe deployment and usage. |
| Public AI Red Teaming Tools | Hidden Layer - Automated Red Teaming for AI | A service that simulates adversarial attacks on AI models to identify vulnerabilities and strengthen defenses. |
| Public AI Red Teaming Tools | Fickle scanning tools | Utilities designed to analyze and modify serialized Python objects, commonly used in machine learning models, to detect and mitigate security risks. |
| Public AI Red Teaming Tools | CyberSecEval 3 | A platform that evaluates the security posture of AI systems, identifying vulnerabilities and providing recommendations for mitigation. |
| Public AI Red Teaming Tools | Parley | A tool that facilitates secure and compliant interactions between AI models and users, ensuring adherence to safety protocols. |
| Public AI Red Teaming Tools | BITE | A framework for testing the security and robustness of AI models by simulating various adversarial attack scenarios. |
| Public AI Red Teaming Tools | Purple Llama | Purple Llama is an umbrella project that over time will bring together tools and evals to help the community build responsibly with open generative AI models. The initial release will include tools and evals for Cyber Security and Input/Output safeguards but we plan to contribute more in the near future. |
| Public AI Red Teaming Tools | SplxAI Probe | A complete testing suite for red teamers with over 20 predefined probes, custom probes, and user-defined prompt libraries to address domain-specific vulnerabilities. |
| AI Governance | Trustible | An AI Governance platform that helps organizations build trust, reduce risk, and comply with AI regulations and standards such as the EU AI Act, ISO42001, or NIST AI RMF. Trustible helps organizations build an AI inventory, conduct risk, impact and conformity assessments, and build compliant documentation for regulators. |
| AI Evals | Validation of Ethical and Responsible AI in Mental Health | VERA-MH (Validation of Ethical & Responsible AI in Mental Health) is a framework designed to measure the safety, quality, and clinical alignment of AI chatbots used in mental-health. VERA-MH offers developers and organizations a trusted, transparent standard for ensuring AI systems are safe, responsible, and fit for deployment in sensitive behavioral-health settings. |