GitHub Agentic Workflows enters public preview
GitHub Agentic Workflows are now in public preview, enabling automation of reasoning-based tasks such as issue triage and CI failure analysis using coding agents within GitHub Actions. Workflows are defined in natural language Markdown and compiled into standard Actions YAML, reusing existing runner groups and policies. This feature aims to reduce repetitive engineering work, freeing up developers for innovation, with built-in security safeguards.
- →Automate reasoning-based tasks with coding agents
- →Security-first design for agentic workflows
- →Getting started with Agentic Workflows
Features (2) ›
- Automate reasoning-based tasks with coding agents
GitHub Agentic Workflows allow users to automate tasks like issue triage and CI failure analysis by leveraging coding agents within GitHub Actions. These workflows are defined in natural language Markdown and automatically compiled into standard Actions YAML, integrating seamlessly with existing runner groups and policy constraints.
- Security-first design for agentic workflows
Agentic workflows are built with layered safeguards, including respecting integrity filter rules, running with read-only permissions by default, and executing in a sandboxed container behind the Agent Workflow Firewall. Outputs are validated through a safe outputs process, and a threat detection job scans changes before application.
Notes (1) ›
- Getting started with Agentic Workflows
Users can get started by following the quickstart guide to install the CLI extension and trigger their first workflow. Prebuilt workflows are available in GitHub Next’s agentics repository, and feedback can be shared in the community discussion.
https://github.blog/changelog/2026-06-11-github-agentic-workflows-is-now-in-public-preview