Sourcegraph Agentic Batch Changes has been made publicly available, allowing enterprise engineering teams to leverage an AI-powered approach to managing mass code changes across thousands of repositories. With the new feature, Sourcegraph combines its Batch Changes execution engine with the power of code search and AI orchestration. In turn, enterprises get an ability to plan, execute, and track mass code changes through a simple prompt with full visibility.
The company launched the feature in order to solve the increasing complexity that engineering teams face when working with today’s software environment. In particular, instead of repetitive manual efforts, teams will be able to automate mass code changes. While still keeping engineers engaged in the approval process. It means that organizations will be able to improve the speed of their software deliveries while maintaining control and consistency.
As Sourcegraph claims, Agentic Batch Changes allows solving one of the most challenging problems in the engineering space. Large organizations often need to change dependencies, update API’s, address security vulnerabilities. And modernize their infrastructure across hundreds or even thousands of repositories. Still, executing such changes consistently often takes considerable manual effort or fragile
“Agentic Batch Changes brings the capabilities of the best coding agents to the largest companies in the world,” said Dan Adler, CEO of Sourcegraph.
AI Simplifies Enterprise-Wide Code Changes
Engineering teams frequently face organization-wide updates that require thorough implementation. Indeed, such updates as dependency upgrades, security patches, API migration, or infrastructure improvements have to be performed for each impacted repository without causing inconsistency. Despite this, scaling these updates has always been a challenging task for companies.
Earlier, there were usually four ways to handle the updates. Teams either manually implemented updates, developed their own scripts, or used coding agents. To perform updates in repositories separately or just skipped large-scale implementation at all. Every approach added extra risks or required additional engineering work.
Using Sourcegraph Agentic Batch Changes, engineers just need to describe the needed update in one prompt. Then, using Sourcegraph Search and Deep Search, the AI agent finds out the impacted repositories. Further, it implements the changes within one repository and gradually rolls out to other repositories. Moreover, the solution adapts to differences between the repositories, receives CI feedback, and retries failing changes before rolling out the updates at scale.
Notably, engineers review and approve every generated changeset before merging.
”Agentic Batch Changes brings the capabilities of the best coding agents to the largest companies in the world,” said Dan Adler, CEO of Sourcegraph. “For years, the owners of the largest codebases have watched smaller competitors move faster because large scale changes were too slow and too risky. Agentic Batch Changes changes that, enabling engineering teams to roll out changes across thousands of repositories and directories with the speed and confidence of changing just one.”
Sourcegraph Agentic Batch Changes Useful for Improving Security on Mercari
In the experimental preview, Mercari implemented Sourcegraph Agentic Batch Changes with a view of improving its security measures. More specifically, Patrick Klitzke, team lead at Mercari, used an AI agent to identify a vulnerability pertaining to GitHub Actions environment variable injections.
“I looked at a GitHub injection issue where you have to set environment variables correctly. I was able to fix it with one prompt on both the Help Center frontend and backend, then extended this to all repos in Mercari – I found around 80 potential repos affected,” Klitzke said.
This shows how the system takes into consideration the variations of the repositories, yet guarantees uniformity of the execution process.
As stated by Klitzke, it should be emphasized that the AI agent performs analysis of each repository individually, without resorting to automation only.
“With the help of Agentic Batch Changes, you’re able to handle repos that have similar, but not identical setups. A normal scripted change would most likely be a text search and replace operation without any context of how it’s actually used,” Klitzke said.
Built for Complex Software Engineering Workflows
Agentic Batch Changes from Sourcegraph was specifically created to provide functionality that surpasses the scope of ordinary scripts in software engineering. Some of the features of the tool include dependency upgrade with breaking change, CVE patching, code pattern refactor, replacement of deprecated APIs, library installation, authentication update, service mesh migration, and CI pipeline modernization.
Moreover, Agentic Batch Changes uses the code graph, Deep Search, and Batch Changes architecture of Sourcegrap. To find out the best execution technique for all projects. While mechanical updates are executed by the help of deterministic scripts. The more complex context-sensitive updates are assigned to sophisticated coding agents like Claude Code and Codex. Instructions for each particular repository are provided through Sourcegraph MCP.
For related updates on digital trust and cybersecurity, explore our SOC News.
Source: Businesswire
