Today's AI Agent: Fastest-Growing Projects — April 26, 2026
Today's AI Agent landscape is marked by a surge in interest around tools that enable humans to work more effectively alongside autonomous agents, as well as those that provide insight into the inner workings of popular coding agents. The top-gaining repositories reflect this trend, with many showcasing innovative approaches to human-agent collaboration and transparency. Meanwhile, projects focused on penetration testing and AI-powered trading agents continue to attract significant attention.
HKUDS/Vibe-Trading, boasting a growth score of 96.80 and 2,614 stars, is an AI-powered trading agent designed to assist users in making informed investment decisions. Its remarkable growth can be attributed to its user-friendly interface and the increasing demand for automated trading solutions.
sooryathejas/METATRON, with a growth score of 89.23 and 2,573 stars, offers an AI-powered penetration testing assistant that leverages local language models on Linux (Parrot OS). Its popularity stems from its ability to streamline the testing process, making it more efficient for security professionals.
wanman by chekusu has gained significant traction with a growth score of 80.75 and 286 stars, as it introduces an open-source agent matrix runtime inspired by Japanese one-man trains. This innovative approach allows humans to step back into an observer role while local agent runtimes coordinate autonomous multi-agent workflows.
by-scott/cortex, sporting a growth score of 79.12 and 39 stars, is a cognitive runtime for language models that features memory, metacognition, multimodal channels, native plugins, and a self-evolving Executive. Its rapid growth can be attributed to its versatility and potential applications in various AI domains.
iamzhihuix/skills-manage has achieved a growth score of 76.50 and 1,104 stars by providing a desktop app that simplifies the management of AI coding agent skills across multiple platforms. This convenience factor has resonated with users seeking to streamline their workflow.
ZeroZ-lab/cc-design boasts a growth score of 75.93 and 631 stars for its high-fidelity HTML design and prototype guidance skill for AI agents. Its popularity can be attributed to the increasing demand for human-AI collaboration tools that enhance design workflows.
Windy3f3f3f3f's how-claude-code-works repository has gained significant attention with a growth score of 74.98 and 2,006 stars by providing an in-depth analysis of Claude Code internals. This transparency has sparked interest among developers seeking to understand the inner workings of popular coding agents.
alejandrobalderas/claude-code-from-source has achieved a growth score of 71.06 and 1,786 stars by reverse-engineering Anthropic's AI coding agent from source maps. Its popularity stems from its ability to provide insight into the architecture and patterns used in Claude Code.
joeynyc/hermes-hudui rounds out our list with a growth score of 65.82 and 1,209 stars for its web UI consciousness monitor designed specifically for Hermes – an AI agent with persistent memory. This innovative approach has piqued the interest of developers exploring human-AI collaboration possibilities.
HKUDS/Vibe-Trading, boasting a growth score of 96.80 and 2,614 stars, is an AI-powered trading agent designed to assist users in making informed investment decisions. Its remarkable growth can be attributed to its user-friendly interface and the increasing demand for automated trading solutions.
sooryathejas/METATRON, with a growth score of 89.23 and 2,573 stars, offers an AI-powered penetration testing assistant that leverages local language models on Linux (Parrot OS). Its popularity stems from its ability to streamline the testing process, making it more efficient for security professionals.
wanman by chekusu has gained significant traction with a growth score of 80.75 and 286 stars, as it introduces an open-source agent matrix runtime inspired by Japanese one-man trains. This innovative approach allows humans to step back into an observer role while local agent runtimes coordinate autonomous multi-agent workflows.
by-scott/cortex, sporting a growth score of 79.12 and 39 stars, is a cognitive runtime for language models that features memory, metacognition, multimodal channels, native plugins, and a self-evolving Executive. Its rapid growth can be attributed to its versatility and potential applications in various AI domains.
iamzhihuix/skills-manage has achieved a growth score of 76.50 and 1,104 stars by providing a desktop app that simplifies the management of AI coding agent skills across multiple platforms. This convenience factor has resonated with users seeking to streamline their workflow.
ZeroZ-lab/cc-design boasts a growth score of 75.93 and 631 stars for its high-fidelity HTML design and prototype guidance skill for AI agents. Its popularity can be attributed to the increasing demand for human-AI collaboration tools that enhance design workflows.
Windy3f3f3f3f's how-claude-code-works repository has gained significant attention with a growth score of 74.98 and 2,006 stars by providing an in-depth analysis of Claude Code internals. This transparency has sparked interest among developers seeking to understand the inner workings of popular coding agents.
alejandrobalderas/claude-code-from-source has achieved a growth score of 71.06 and 1,786 stars by reverse-engineering Anthropic's AI coding agent from source maps. Its popularity stems from its ability to provide insight into the architecture and patterns used in Claude Code.
joeynyc/hermes-hudui rounds out our list with a growth score of 65.82 and 1,209 stars for its web UI consciousness monitor designed specifically for Hermes – an AI agent with persistent memory. This innovative approach has piqued the interest of developers exploring human-AI collaboration possibilities.