Daily Papers of 2026-08-19

  1. Demystifying Agent Skills: Why They Work-Until They Don't 165 upvotes, #1 of 2026-08-19
  2. FACET: Preserving Source Intent and Executable State in Terminal Task Synthesis 119 upvotes, #2 of 2026-08-19
  3. Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements 105 upvotes, #3 of 2026-08-19
  4. FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution 97 upvotes, #4 of 2026-08-19
  5. ASI-Bench: At the Dawn of Artificial Superintelligence 62 upvotes, #5 of 2026-08-19
  6. Embodied-Navigator: Point, Think, Memorize, and Align for Efficient Navigation 51 upvotes, #6 of 2026-08-19
  7. Agent Lightning v1.0: Towards Harnessed Agentic RL 31 upvotes, #7 of 2026-08-19
  8. V-RAE: Rethinking Video Latent Spaces for Generation 30 upvotes, #8 of 2026-08-19
  9. LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents 24 upvotes, #9 of 2026-08-19
  10. EDITBRIDGE: Towards Faithful and Efficient Ultra-High-Resolution Image Editing 22 upvotes, #10 of 2026-08-19
  11. DiSCO: Defending text-to-image generation through distribution-guided contrastive prompt optimization 20 upvotes, #11 of 2026-08-19
  12. Dynamic Multi-Byte Prediction With Hierarchical Language Models 19 upvotes, #12 of 2026-08-19
  13. MoE-ViE: Mixture of Experts Vision Encoder for Efficient Image and Video Understanding 16 upvotes, #13 of 2026-08-19
  14. Harness the Memory: A Holistic Evaluation of Memory Substrates in Memory Agents 15 upvotes, #14 of 2026-08-19
  15. CoinVE-200K: A Large-Scale High-Quality Dataset for Compositional Instruction-Guided Video Editing 15 upvotes, #14 of 2026-08-19
  16. From Corpora to Co-Evolving Capabilities: Capability-Centric Data Design for Generalist Image Generation 12 upvotes, #16 of 2026-08-19
  17. Energy-Guided Flow Matching 11 upvotes, #17 of 2026-08-19
  18. SkillForge: Self-Distilling Agents for Project-Specific Issue Resolution 11 upvotes, #17 of 2026-08-19
  19. Abra: Scaling Diffusion Image Training 10 upvotes, #19 of 2026-08-19
  20. HarnessRisk: A Lifecycle-Oriented Benchmark for Agent Harness Safety 10 upvotes, #19 of 2026-08-19
  21. StartupBench: Benchmarking General-Purpose Agents on Market-Validated End-to-End Workflows 9 upvotes, #21 of 2026-08-19
  22. MathForm: Scaling Mathematical Autoformalization with Knowledge Retrieval and Verification-Guided Refinement 8 upvotes, #22 of 2026-08-19
  23. From Sequence to Structure: Relational Uncertainty Propagation for LLM Agents 7 upvotes, #23 of 2026-08-19
  24. Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt Injection 6 upvotes, #24 of 2026-08-19
  25. Cross-Model Memory Transfer via Target-Side Reader Adaptation 6 upvotes, #24 of 2026-08-19
  26. Personalized Auto-Research: Towards a True AI Co-Scientist 5 upvotes, #26 of 2026-08-19
  27. Unifying Graph Neural Networks Through a Common Layer Equation 5 upvotes, #26 of 2026-08-19
  28. aDSL: Agentic 3D Creation via Joint Agent-Program Design 5 upvotes, #26 of 2026-08-19
  29. The Problem Is the Problem: Towards Scalable Mathematical Discovery 4 upvotes, #29 of 2026-08-19
  30. GS-Voxel: Fitting-Free Structured Latents for Large-Scale 3DGS Generation 4 upvotes, #29 of 2026-08-19
  31. PTXBench: Benchmark and Adapt LLMs for GPU Kernel Optimization with Architecture-specific PTX 3 upvotes, #31 of 2026-08-19
  32. CardioState-JEPA: Delay-Aware Cross-Modal Learning of a Shared Cardiac Representation 2 upvotes, #32 of 2026-08-19
  33. PixRestore: Unified Image Restoration via Pixel Diffusion Transformer 2 upvotes, #32 of 2026-08-19

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.