2026年8月20日 技术热点总结

项目简介：A compact Agent Skill for highly simplified, rounded, subtly neo-skeuomorphic IP mascot logos.

技术栈：codex、codex-skill、image-generation、logo-design、mascot-design

核心介绍：ip-as-logo is a compact Agent Skill for generating highly simplified company-ready mascot logos. It treats the result as a logo first and a character second: familiar cute animals by default, bold rounded silhouettes, strict complexity limits, oversized corner composition, and extremely subtle neo-skeuomorphic shading. It follows the open Agent Skills format and is designed to work with any compatible AI agent…

项目简介：The plugin market inside DeepSeek Harness — browse, search, one-click install · DSH 可视化插件市场

技术栈：TypeScript、deepseek-harness、dsh-plugin、marketplace

核心介绍：English | 中文 The plugin market inside DeepSeek Harness. Open Settings → Plugin Market → browse, search, one-click install. One-click themes — install, switch live, no restart:

项目数据：⭐ 1,245 Stars，🍴 77 Forks

项目简介：Buyer-run, ad-neutral shopping-agent MCP software with deterministic ranking, signed purchase mandates, and a local audit trail.

技术栈：JavaScript、agentic-commerce、human-in-the-loop、local-first、mcp、mcp-server、model-context-protocol、privacy、self-hosted

核心介绍：NorthCinder is an open-source MCP server that helps an AI agent compare products against your brief. It returns a ranked shortlist with reasons, reports which stores it could and could not search, and requires a separate approval before checkout. Seller payment and affiliate data never improve a result’s position. NorthCin…

项目简介：DeepSeek Harness Desktop

技术栈：TypeScript、ai-agents、deepseek、desktop、dsh-plugins

核心介绍：A local-first, cross-platform desktop shell for DeepSeek Harness. English · 简体中文 Beyond official DeepSeek models, DSH Desktop supports mainstream third-party model providers—with more DSH-powered desktop experiences coming soon.

项目数据：⭐ 1,096 Stars，🍴 94 Forks

AI systems are increasingly capable of contributing to mathematical research. In research practice, frontier-model reasoning is a limited resource, and expert mathematical review is even more sharply constrained. Allocating these scarce resources well is therefore central to making AI-assisted mathematical discovery efficient. In most current AI-for-math workflows, human effort is concentrated at the beginning and end, in selecting suitable research problems and later reviewing the resulting artifacts. These two stages are becoming bottlenecks for research-level mathematics. We address them…

Reinforcement learning for coding agents increasingly relies on long-running agent harnesses to manage tool integration, repository contexts, and execution feedback. However, the native execution environments of these harnesses are inherently misaligned with policy-gradient training: environmental crashes and reward hacking corrupt outcome signals, while train-inference discrepancies decouple rollout behavior from policy updates. To address this, we present LEGO-RL, a framework that bridges native coding-agent harnesses with scalable policy-gradient optimization without modifying their inte…

We introduce PTXBench, a benchmark for evaluating and adapting large language models (LLMs) to use architecture-specific PTX for GPU kernel optimization. PTXBench measures functional correctness, whether selected target instructions execute at runtime, and speedup over frontier libraries across GEMM and attention workloads on H100 and B200 GPUs. Our evaluation shows that architecture-specific PTX capability remains uneven: success rates fall substantially on complex attention backward workloads, and executing the target instructions does not necessarily translate into competitive performanc…

Electrocardiography (ECG), photoplethysmography (PPG), and phonocardiography (PCG) provide complementary views of the same cardiac cycle, yet existing cardiac foundation models are trained for a single sensing modality, leaving the shared physiology across sensors unexploited. We introduce CardioState-JEPA, a cardiac foundation model to learn a single shared representation jointly across ECG, PPG, and PCG, built on a physiology-aware joint-embedding predictive architecture. The model maps heterogeneous waveforms into a common token space, processes them with a single shared Transformer enco…

Latent video generation relies on autoencoders to define a compact space in which generative models operate. Although video autoencoder architectures have evolved substantially, their latent spaces are still optimized primarily for pixel-level reconstruction and provide limited high-level semantic organization. A reconstruction-optimal latent space, however, need not be well suited to generative modeling. We propose V-RAE, a video representation autoencoder that builds compact generative latents on top of frozen vision foundation model representations. A lightweight temporal pooling module …

As text-to-image generative models advance, they raise critical safety concerns, particularly the generation of Not-Safe-For-Work (NSFW) content such as violence and nudity, further exacerbated by red-teaming adversarial attacks. Existing defenses predominantly operate under white-box assumptions, relying on text encoder optimization, weight editing, or inference-time intervention, and fundamentally cannot scale to proprietary models. Black-box alternatives based on LLM prompt rewriting offer broader applicability, yet fail in a critical regime we identify as the benign adversarial problem:…