{"data":{"id":"36dd7c0f-a1df-4dd1-b8b0-3fd444afa980","title":"VAST: Compiling Grounded Vulnerability Scenarios into Verifiable, Multi-Step Environments for Evaluating LLM Red-Team Agents","summary":"VAST is a declarative range compiler that builds reproducible, verifiable multi-step environments from vulnerability scenarios to evaluate LLM red-team agents. In a budget-matched prompt contrast on the TopicServer stratum, three models solved 12 to 14 of 90 instances under contract-only or generic-hint prompts, rising to 76 of 90 when the shared workflow protocol was supplied in full.","solution":"N/A -- no mitigation discussed in source.","labels":["research","security"],"sourceUrl":"https://arxiv.org/abs/2611.19465v1","publishedAt":"2026-11-14T15:40:08.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"info","attackType":[],"issueType":"research","affectedPackages":null,"affectedPackageNames":null,"affectedPackageRefs":null,"affectedVendors":[],"affectedVendorsRaw":["glm-5.3","grok-4.7","deepseek-v4-pro","Qwen2.5-Coder-7B-Instruct"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","summaryPromptVersion":"v2","headline":null,"headlinePromptVersion":null,"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"affectedPackagesCheckedAt":null,"patchAvailable":null,"disclosureDate":"2026-11-14T15:40:08.000Z","fixReleasedAt":null,"daysToFix":null,"fixCheckedAt":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":null,"aiComponentTargeted":"agent","llmSpecific":true,"classifierConfidence":0.9,"researchCategory":"preprint","atlasIds":null}}