Claude Opus 4.8 on GitLab: Complex Agentic Work, Less Disruption
Anthropic’s Claude Opus 4.8 has been added to GitLab’s Duo Agent Platform, targeting complex, multi‑step agentic workflows. The model promises tighter execution, fewer manual corrections, and stronger performance on coding, document drafting, and data analysis. It also introduces mid‑conversation system prompts that can be updated without flushing the prompt cache. Opus 4.8 runs on GitLab Credits and is available immediately to free‑trial, Premium and Ultimate users.
Full Security Scanner Coverage of Your Codebase in Minutes
GitLab 19.0 introduces security configuration profiles that centralize SAST, dependency scanning, and secret detection settings. Teams can apply these default profiles across any number of projects from the UI, eliminating the need to edit individual .gitlab-ci.yml files. The profiles automatically...
More AI Models for GitLab Duo Agent Platform Self-Hosted
GitLab 19.0 expands its Duo Agent Platform Self‑Hosted offering by supporting four new open‑source large language models—Mistral Devstral 2 123B, GLM‑5.1, Kimi‑K2.6 and MiniMax‑M2.7. The models run on‑premises via the vLLM serving platform, enabling fully air‑gapped or hybrid deployments without sending code...
Transform MRs From Manual Tasks to an Automated Workflow
GitLab 19.0 introduces Developer Flow, an AI‑driven agent that automates the entire merge request (MR) lifecycle—from generating an MR to handling reviewer feedback, resolving conflicts, and performing a one‑click rebase and merge. The agent can be triggered via a button,...
Track CI Component Usage Across Your Organization
GitLab 19.0 introduces a Components Analytics view in the CI/CD Catalog, giving platform teams visibility into how shared pipeline components are adopted across the organization. A high‑level usage dashboard is available to all tiers, while Ultimate customers can drill down...
Beyond BYOK: Why Governance Matters for AI Agents
GitHub announced that its Copilot CLI now supports bring‑your‑own‑key (BYOK) and fully offline models, letting developers route requests to private providers or run models locally. In contrast, GitLab’s Duo CLI is built on the Duo Agent Platform and provides headless,...
GitLab Dedicated for Government Now GovRAMP-Authorized
GitLab announced that its Dedicated for Government offering has earned GovRAMP authorization, clearing a major procurement hurdle for state and local agencies. The single‑tenant SaaS solution provides U.S.‑only data residency, isolated infrastructure, and private networking to satisfy stringent compliance rules....
5 Ways to Fix Misleading Vulnerability Severities with Policy
GitLab introduced vulnerability management severity override policies that let teams automatically adjust CVSS scores based on context such as CVE ID, CWE, file path, or exploitation intelligence. The policies can set, increase, or decrease severity levels on each default‑branch pipeline,...
Consolidate Your GitLab Stack with Gitaly on Kubernetes
GitLab 18.11 makes Gitaly on Kubernetes generally available, letting teams run the entire GitLab stack inside a single Kubernetes cluster. Previously, organizations had to maintain a hybrid architecture with Gitaly on virtual machines, adding operational complexity. The new deployment uses...
Limit Credential Exposure with Fine-Grained Personal Access Tokens
GitLab has launched a beta for fine‑grained personal access tokens (PATs), allowing users to restrict a token’s reach to selected projects or groups and assign per‑resource permissions such as Create, Read, Update, and Delete. This replaces the traditional broad‑scoped tokens...
8 Agentic AI Patterns Reshaping Team Collaboration
The article maps eight recurring AI‑agent patterns that are reshaping how teams collaborate, from automated status updates to role‑specific agents embedded in chat. These patterns consistently deliver three outcomes—moving faster, working smarter, and staying in control—by reducing manual coordination, improving...
Atlassian Will Train on Your Data: Opt Out with GitLab
Atlassian will automatically collect metadata and in‑app content from its cloud suite—including Jira, Confluence, Bitbucket and related products—starting August 17 2026 to train AI models like Rovo. The collection is mandatory for Free, Standard and Premium tiers, with only Enterprise customers able...
How to Detect and Prevent Contagious Interview IDE Attacks
GitLab’s Threat Intelligence team uncovered a North Korean “Contagious Interview” campaign that leverages malicious Visual Studio Code tasks to execute malware when a repository is opened. By analyzing the VS Code source, GitLab pinpointed the node‑pty “spawn‑helper” binary as a low‑level...
Teaching Software Development the Easy Way Using GitLab
University of Washington‑Bothell lecturer Stephen G. Dame shows how GitLab for Education lets instructors distribute assignments, manage permissions, and deliver feedback at scale. By leveraging GitLab Ultimate’s Groups, Subgroups and merge‑request workflow, educators can mirror real‑world development environments while keeping...
GitLab and Anthropic: Governed AI for Enterprise Development
GitLab has deepened its integration with Anthropic’s Claude model, making it the default AI engine within the GitLab Duo Agent Platform. The partnership embeds Claude‑driven code generation, review, chat, and vulnerability resolution directly into GitLab’s existing software development lifecycle, preserving...