2026年7月25日 技术热点总结

项目简介：A tactile WebGL sticker maker with rich text, image uploads, and interactive peel physics.

技术栈：JavaScript、sticker、webgl

核心介绍：Sticker Forge turns text or uploaded image artwork into a tactile WebGL sticker. Grab the real die-cut edge, drag it inward, and the sticker curls to reveal a satin back surface with depth-aware shadowing. The repository contains two deliverables: Node.js 22.13 or newer is required. Build the hosted app and both reusable library bundles with: The reusable files are emitted to public/embed/: Pushes to main deploy a static export through the GitHub Pa…

项目简介：Grok free-register CLI: register → OAuth → CPA JSON (Go)

技术栈：Go

核心介绍：Grok 免费号 注册 → OAuth → CPA 可用 JSON 二合一 CLI（Go）。 一条命令后台跑完，产物可直接导入 CPA / cliproxy 类网关。 grok start # 交互：数量 + 线程(1–8) grok start -t 10 –thread 2 grok config # 编辑 ~/.grok/config.env grok upload # 手动上传 CPA JSON 到 Management API 协议腿 gRPC 发/验码 → Server Action → SSO hop → OAuth → 探活/CPA 边缘腿 chrome_131 TLS 优先 → CF 拦再 clearance（auto） 挑战腿 Turnstile 仅 Chromium（offscreen 池 / one-shot） 冒烟（不注册账号）: go run scripts/smoke_protocol.go

项目数据：⭐ 424 Stars，🍴 147 Forks

项目简介：the optimized, token-efficient version of the leaked Claude Fable 5 / Mythos 5 system prompt. Re-engineered into clean Markdown for universal execution on Gemini 3.1 Pro, ChatGPT 5.6, and advanced LLM agents.

核心介绍：A vendor-neutral adaptation of Anthropic’s publicly available Claude Fable 5 / Mythos 5 system prompt. The original prompt contains a large amount of infrastructure that exists specifically for Claude’s production environment, including XML conventions, tool interfaces, product-specific behavior, and internal implementation guidance. While appropriate for Claude,…

技术栈：Python

核心介绍：𝕏 slvDev · LinkedIn This is a 28.9 million parameter language model that generates text on an ESP32-S3, a microcontroller that costs about $8. It runs on the chip itself, with nothing sent to a server, and it writes each word to a small screen wired to the chip at roughly 9 tokens per second. The last language model people ran on a chip like this had 260 thousand parameters, so this one holds about a hundred times more. It fits because

项目数据：⭐ 387 Stars，🍴 52 Forks

项目简介：RCE PoC for Redis 6.2.22, 7.4.9, 8.6.4, 8.8.0

技术栈：Python

核心介绍：Non-destructive RCE exploits for Redis 6.2.22, 7.4.9, 8.6.4 via the stream consumer-group shared-NACK double free (a patch bypass of CVE-2026-25243), 8.8.0 via a TDigest heap-overflow in bundled RedisBloom, and 8.8.1 via a TopK wild free (a patch bypass of TopK wild free; bypass of the incomplete CVE-2026-25589 fix — works on RedisBloom v8.8.0 and v8.8.2) needed to build valid RESTORE payloads) gcc -shared -fPIC -O2 -o libcrc64.so crc64.c crcspeed.c

项目数据：⭐ 385 Stars，🍴 99 Forks

Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also make agents hard to train end-to-end with open infrastructure, whose SFT/RL stacks cannot natively express stateful, multi-process harness inference. To address this, we present OpenForgeRL, an open-source framework for training harness-based agents end-to-end in diverse environments. OpenForgeRL achieves this with a lightweight proxy that serves the harness's model calls while re…

We revisit dataset distillation from an outcome-centric perspective. Rather than aligning process surrogates (per-step gradients or training trajectories), Influence Matching (Inf-Match) aligns the final outcome of training: it learns a compact synthetic set whose effect on the converged parameters matches that of the full dataset. Concretely, we introduce a fully differentiable, sample-level influence estimator that quantifies parameter shifts from adding or removing data, without time-consuming inverse-Hessian products or convexity assumptions. The estimator runs in linear time by unrolli…

We study on-policy distillation (OPD) for agentic tasks, where an LLM agent interacts with an environment over multiple turns and a student imitates a teacher over these multi-turn interaction histories. Fully online OPD is costly because each update requires fresh student rollouts through the environment and teacher queries at visited histories. We propose Replayed-Prefix On-Policy Distillation (ReOPD), an off-environment alternative that reuses pre-collected teacher trajectories as replayed prefixes: the student acts at selected steps, while the teacher provides dense per-step supervision…

We introduce SANA-Video 2.0, a hybrid video diffusion transformer instantiated at 5B and 14B scales under a unified architecture. Designed to generate high-quality video up to 720p on a single GPU, SANA-Video 2.0 matches full-softmax video DiTs in quality while retaining the favorable long-sequence scaling of linear attention. To avoid quadratic attention throughout, Hybrid Linear-Softmax Attention combines gated linear attention for O(N)-dominated mixing with periodic gated-softmax anchors at a 3:1 ratio, restoring the full-rank token interactions that pure linear attention lacks. To propa…

Agentic Reasoning has become a transformative force in financial analysis due to its ability to integrate large-scale information and generate reliable and accurate content. However, when handling complex real-world problems, different agents still show significant performance variation. In this work, we design Finance-LaTeX SKILL, a skill for synthesizing financial documents with complex layouts based on expert knowledge. Using an agent workflow built on this skill, we generate 2,000 professional financial documents along with 6,000 high-quality question-answer pairs. To evaluate the overa…

Understanding motion in video is a fundamental challenge for visual learning, as frame-to-frame change entangles two sources of dynamics: camera motion and object motion. This decomposition has remained underexplored in representation learning, partly because these factors are tightly coupled in natural videos and difficult to supervise separately. Yet recovering it is important for learning robust motion representations that separate meaningful object dynamics from camera-induced variation. We study whether such structured motion representations can be recovered from frozen features of a p…