https://cybertoolchain.io/tools/vega New releases and features in Vega, tracked by The Cyber Toolchain. en Tue, 04 Aug 2026 00:00:00 GMT https://vega.io/blog/vega-introduces-detection-skills https://vega.io/blog/vega-introduces-detection-skills Tue, 04 Aug 2026 00:00:00 GMT Vega launches Detection Skills, an open standard for post-match triage, investigation, and optimization built on Anthropic's Agent Skills format. • Launches the Detection Skills open standard at detectionskills.io, built on Anthropic's Agent Skills format, for encoding triage, investigation, and detection-optimization logic as versioned, shareable YAML skills. • Integrates Detection Skills into the Vega Cyber Defense Platform via a detection notebook workflow: compose detection logic, attach skills for triage/investigation/optimization, and receive tuning proposals with attached reasoning for engineer approval. • Adds an optimization loop that writes approved verdicts back into the detection rule, incrementally sharpening coverage with each alert fire. • Enables agentic detection and triage for emerging threat landscapes including dangerous AI model activity, with skills scoped to specific use cases rather than relying on a single catch-all agent. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/vega https://vega.io/blog/vega-introduces-detection-skills https://vega.io/blog/vega-introduces-detection-skills Tue, 04 Aug 2026 00:00:00 GMT Vega introduces Detection Skills: an open standard for attaching triage, investigation, and tuning logic directly to detections. • Launches Detection Skills, an open standard built on Anthropic's Agent Skills format, enabling triage, investigation, and optimization reasoning to be written as versioned skills that attach to detections and execute as a continuous agentic loop. • Ships the Detection Skills Library with 50+ skills from Vega Research and partners — the same skills running inside the Vega platform — available openly at detectionskills.io. • Provides a sandbox at detectionskills.io to build a detection, attach a skill, preview the verdict, and export spec-compliant YAML. • Hosts a GitHub repository for the community to contribute Detection Skills back to the shared library. • Adds a three-step Detection Skills workflow inside the Vega Cyber Defense Platform: compose detections in a detection notebook, attach skills from the Skill Library for triage/investigation/optimization, and approve or reject tuning proposals that write approved verdicts back into the rule. • Enables noise prevention by running triage skills before an alert surfaces, rather than after escalation. • Adds coverage for agentic AI threat scenarios — 'dangerous model activity' — through new Detection Skills targeting that use case. Full history on Cyber Toolchain: https://cybertoolchain.io/tools/vega