GitHub expanded AI-powered secret detection to 37 new

GitHub added 37 new secret detectors in March and extended secret scanning directly into AI coding agents via MCP protocol. The GitHub MCP Server now scans code changes for exposed secrets before commits or pull requests are made, catching credentials from AI-generated code before they enter repositories. Push protection expanded to 39 detectors by default with validity checks for major providers including DeepSeek, npm, and Sentry tokens.

GitHub announced AI-powered security detections to expand application security coverage across more languages and frameworks beyond traditional CodeQL support. The new detections complement static analysis by surfacing vulnerabilities in areas difficult for traditional tools. Public preview planned for early Q2 2026.

GitLab's agentic AI platform is now generally available, providing orchestration across the entire software development lifecycle. The platform includes seven AI agents for tasks from planning to security, with multi-step reasoning capabilities and support for multiple AI models. Features central AI Catalog for agent management and customizable flows for automated workflows.

Atlassian announced OAuth 2.0 as a new authentication type for Cloud Application Links in Jira and Confluence, offering industry-standard secure authentication protocol. The update also includes AI-powered false positive detection for secret scanning that analyzes findings before developers see them, reducing security alert fatigue.

GitGuardian released ML-powered similar incident grouping to combat security alert fatigue by automatically clustering related security incidents for efficient bulk remediation. The feature analyzes incident context beyond detector types to identify meaningful relationships, reducing manual security triage work.

New analysis shows significant outage patterns across DevOps platforms in 2024: Azure DevOps had over 110 incidents totaling 826 hours downtime, GitHub recorded 124 incidents with 18 vulnerabilities patched, and Atlassian platforms saw 30% increase in incidents. The shared responsibility model leaves organizations ultimately responsible for data protection during outages.