Multi-Sourced, Multi-Agent Evidence Retrieval for Fact-Checking Jailbreak Foundry: From Papers to Runnable Attacks for Reproducible Benchmarking From Variance to Invariance: Qualitative Content Analysis for Narrative Graph Annotation Real Money, Fake Models: Deceptive Model Claims in Shadow APIs Pseudo Contrastive Learning for Diagram Comprehension in Multimodal Models