GitLab 18.7: Advancing AI automation and developer experience

What this post added

Introduced Custom Flows for automating multistep workflows using YAML-defined sequences that orchestrate agents. Added a SAST False Positive Detection Flow that uses AI to identify potential false positives. Implemented Custom Agent Versioning to allow teams to pin specific versions of AI Catalog agents and flows. Enabled administrators to turn Foundational Agents on or off at the instance or group level. Launched the Data Analyst Agent for natural language exploration of GitLab data using GLQL. Added Dynamic Input Selection to GitLab Pipelines for UI-driven pipeline triggering with dynamic, cascading dropdown fields. Introduced CI/CD Catalog Publication Guardrails for administrators to control which projects can publish components. Added Warn Mode for MR Approval Policies, allowing violations to be surfaced without blocking merges, and displaying policy violations in the Vulnerability Report with a visual badge.