Autoresearch (karpathy): The Dawn of Autonomous AI Research Swarms
Discover how Andrej Karpathy's autoresearch uses autonomous AI agents to iterate on LLM training code overnight for continuous research progress.
Stop Wasting Tokens: How Ponytail 5 Transforms AI Code Generation
Reduce code by 53% and increase test coverage to 98% with Ponytail 5—the specialized prompt for high-performance AI agents.
REA: Empowering AI Agents to Reverse Engineer Anything
Automate reverse engineering workflows. Use REA to let AI agents inspect binaries, JavaScript apps, and websites locally using MCP.
Automating the Scientific Method: An Introduction to OpenResearch
Transform coding agents into research powerhouses with OpenResearch. A local-first, MIT-licensed workspace by alphaXiv for iterative scientific discovery.
Securing Your AI: A Guide to DeepTeam LLM Red Teaming
Secure your AI agents and RAG pipelines with DeepTeam, an open-source framework for simulating jailbreaking, prompt injection, and more.
Moving Beyond Trial-and-Error: Mastering LLM Evaluation with Promptfoo
Stop using trial-and-error for LLMs. Learn how to use Promptfoo to automate evaluations, run red teaming tests, and secure your AI applications.
Proactive AI Security: Mastering Red Teaming with Microsoft PyRIT
Learn how to proactively identify risks in generative AI systems using Microsoft's open-source Python Risk Identification Tool (PyRIT).
Securing Generative AI: A Deep Dive into garak LLM Vulnerability Scanner
Protect your AI applications from prompt injection and jailbreaks using garak, NVIDIA's open-source LLM vulnerability scanner.
Atomic Agent v0.6.5: Mastering Parallelism with the Fusion Model
Discover how Atomic Agent v0.6.5 uses the new Fusion model to orchestrate parallel workers, enabling efficient local and cloud-based AI workflows.
Accelerating CI/CD: A Guide to Open-Source AI Code Reviews with PR-Agent
Streamline code reviews and reduce costs using PR-Agent, the open-source AI agent for GitHub, GitLab, and more. Full control over your data and models.