Harness Unfurls Source Code Repository Alternative to GitHub

Harness recently unveiled an alternative to GitHub's source code repository service, specifically tailored for DevOps teams that utilize artificial intelligence agents to generate code. Field CTO Martin Reynolds explained that the Agent-Ready Harness Code Repository and AI Code Review service caters to the unique demands of AI agents operating at machine speed.

Unlike traditional repositories, Harness's service can handle thousands of simultaneous pull requests and commits from both AI agents and human developers without compromising search, history, or diffs. The platform is compatible with the Harness Model Context Protocol (MCP) server or command line interface (CLI), allowing DevOps teams to review pull requests by author email, view all open pull requests across repositories, and manage comment threads without a browser.

AI coding agents can be granted specific permissions or inherit them from the teams deploying them. Developers can further define what AI agents can access, merge, or deploy using role-based access controls (RBACs) and policies based on the Open Policy Agent (OPA) framework. The AI Code Review examines pull requests, checks mandatory gates, and prevents any request that fails from being merged.

Differences are grouped by risk, ensuring high-risk changes are not lost in the sea of renamed repositories and dependencies. A feedback capability describes the impact of a change, making it easier to decide whether to merge. Harness Code Repository offers a free trial, with migrations from other repositories, including GitHub, GitLab, Bitbucket, and Azure DevOps, requiring only a few clicks.

As DevOps teams recognize the limitations of a piecemeal approach to agentic AI engineering, Harness suggests that the entire software development lifecycle (SDLC) needs to be re-engineered to manage the increasing amount of code. Harness claims that teams have experienced savings of 10,000 hours in review and approval times since early testing.

Experts agree that accountability remains crucial, regardless of the approach to software engineering in the age of AI.