AI Weekly review · 2026Week 39 · 09.21 — 09.27AI Weekly
NEXGATEHUB AI NEWS
GitHub double release: agent fuzzing and Copilot custom workflows
First backfill edition: covers only previously collected sources; coverage is incomplete. This edition covers September 21 to September 27, 2026. The two most important updates during this period both came from GitHub: first, Antonio Morales of GitHub Security Lab released Fuzzing Taskflow, an autonomous fuzzing pipeline for C/C++ projects built on the Taskflow Agent; users only need to specify a GitHub repository, and the system automatically identifies entry points, analyzes the build system, and writes harnesses, demonstrating LLM agents' real-world application in security testing automation. Second, GitHub officially released a Copilot app tutorial explaining how to run the /create-canvas skill in an agent session, using natural language to describe the workflow, the actions users can take, and the actions the agent can take, thereby generating kanban boards, checklists, dashboards, and other custom workflows. Together, the two updates point to the trend of agent tooling and workflow customization.
LLM agents automate security testing
Section overviewAntonio Morales of GitHub Security Lab released Fuzzing Taskflow, an autonomous fuzzing pipeline for C/C++ projects built on the Taskflow Agent. Users only need to specify a GitHub repository, and the system automatically completes three key tasks: identifying project entry points, analyzing the build system, and writing fuzzing harnesses. It hands the traditionally expert-dependent preliminary fuzzing work over to LLM agents, lowering the barrier to C/C++ security testing.
GitHub Blog · AI & MLFirst-party09.25AI-powered fuzzing with the GitHub Security Lab Taskflow Agent
Antonio Morales of GitHub Security Lab has released Fuzzing Taskflow, an autonomous fuzzing pipeline for C/C++ projects built on the Taskflow Agent — you only specify a GitHub repository, and it automatically identifies entry points, analyzes the build system, writes harnesses, runs AFL++, and triages crashes.
Getting started with Copilot custom workflows
Section overviewGitHub's official tutorial explains how to build custom workflows with canvases in the GitHub Copilot app: run the /create-canvas skill in an agent session, describe the workflow, the actions users can perform, and the actions the agent can perform in natural language, and the agent then generates kanban boards, checklists, dashboards, and other interfaces. This provides an official path for users to solidify repetitive tasks into visual collaboration panels.
GitHub Blog · AI & MLFirst-party09.26GitHub Copilot app for Beginners: How to build custom workflows with canvases
An official GitHub tutorial shows how to build custom workflows with canvases in the GitHub Copilot app: run the /create-canvas skill in an agent session, describe the workflow, user actions, and agent actions in natural language, and the agent generates boards, checklists, dashboards, and other interfaces that open in the side panel.