All repos — GH Trending Explorer

Abliteration without the weights — projective refusal steering as a GLP vector + fail-closed boot hotfix.

Verifiable multi-agent patch evaluation infrastructure

Split your metered internet bundle fairly across every person in the house. Each user gets an allowance (fixed GB, or an equal share of what's left), their devices all share it, and the moment the allowance runs out every device they own is cut at once.

WangDXXXX / budget-sandbox-skill

把预算 Excel 变成可拖拽、可交互、数字精确的「预算沙盘」网页 · Turn a budget Excel into an accurate interactive cash-runway sandbox. Fixed verified engine — AI never rewrites the math.

Reederey87 / glm53-flash-exl3-2x-dgx-spark

GLM-5.3-Flash EXL3 (320B MoE) on 2x NVIDIA DGX Spark — production serving kit, 1M context, 97%+ multi-session prefix caching, DFlash2 spec decode

Git-on-my-level / self-improving-agent-ecosystem

Reference architecture, contracts, skills, and fixtures for safe evaluator-driven self-improving ecosystems.

xyzDist / H3-LongTakeNoCuts

Attempts to fix the H3 motion-context degradation issue

GeekatplayStudio / Meshwright

Know why your model won't print. Fix it. Prove it. — mesh analysis, repair and smart retopology for 3D printing, by Geekatplay Studio

0xBakeer / Qwen3.8-27B-FP8-on-a-single-DGX-Spark

Serving Qwen3.8-27B-FP8 on a single DGX Spark (GB10): 7.88 to 58.5 tok/s single-stream from decode strategy alone, weights untouched. Speculative decoding and prefix caching benchmarked, plus DFlash 2 — the only Qwen3.8-27B build that can serve it under vLLM.

Amduraznak / minimax-h3-fp16-fix

Run MiniMax H3 at native fp16 on pre-bf16 GPUs (V100/Volta) — exact-math fix for the black-frame overflows, ~11× faster than the fp32 fallback.

VinvAI / VinvAI

Tools for AI agents to test, fix and optimise your codebase

ibadwi / arabic-rtl-fixer-ai-skill

AI skill for fixing Arabic RTL, BiDi, and mixed Arabic-English content in documents, Word (DOCX), PowerPoint (PPTX), and other formatted files.

tonyd2wild / MiMo-V2.6-Flash-DGX-Spark-Recipe

MiMo-V2.6-Flash-RL on 2x NVIDIA DGX Spark (GB10): vLLM TP2 + DFlash + vision/audio, fp8 KV 300K, four fixes for the stock image, measured baselines

dharmendradevops11 / the-self-healing-pipeline

Companion code for the book The Self-Healing Pipeline: How I Taught My CI/CD Pipeline to Fix Itself at 2 AM

ziplab / ZipTok3D

ZipTok3D: High-Fidelity 3D Tokenization with Compact Token Prefixes

thecodacus / spec-wins

Three trap-laden fix-it tasks with hidden graders for testing an AI coding agent's judgment. Sandbox setup, prompts, answer key, and results for Qwen3.8-Flash, Qwen3.6-35B and Claude Opus 5.

happy100-dotcom / baidu-autosave-fixed

Automated Baidu Netdisk save tool with file date/size filtering, cross-cloud-drive resource search, scheduled tasks, and WeChat Work notifications. Built with Flask + Vue 3.

Qin-sor / agent-context-lab

A minimal coding-agent harness built to measure what context management actually costs: constraint survival after compaction (8%→83%), default-FAIL verification against false completions, and KV-cache prefix design (21%→90%). Stdlib-only, real deepseek-chat runs

Schaka / rocm-gfx803

ROCm for gfx803 aka RX 460, RX 470, RX 480, RX 550, RX 560, RX 570, RX 580 and RX 590 with custom patches, kernel fixes, etc

alexyc9381 / hiring-panel-skill

Seven Claudes try to reject your resume, then it rebuilds it with every mistake fixed. Free Claude Code skill, MIT.

agent0ai / depfix

Python dependencies, solved.

ptflp / 4deus-mod

Modular Steam Deck fixes and customizations for Decky Loader

MiaAI-Lab / Ornith-1.5-35B-A3B-DGX-Spark

Serve Ornith-1.5-35B-A3B (NVFP4 + in-checkpoint MTP, 256k ctx) on one DGX Spark — vLLM b12x recipe plus CUDA-graph fixes.

StarWorkshop / wavrail

Streaming speech recognition toolkit: log-mel features, CTC greedy and prefix-beam decoding, transducer-style frame-synchronous emission, forced alignment, Chinese/English CER-WER metrics, and a chunked streaming pipeline with latency accounting. NumPy core with a CPU-only torch extra; trainable demos on synthetic audio, fully offline.

YellowPancake / quant-autoresearch

A minimal framework for Agents to research quantitative strategies under a fixed evaluation protocol.

Off-by-1-Labs / FLAWED

Fix-Like Artifacts With Embedded Defects

stamparm / thebleep

Fixes the console command that just failed, or the one you are about to run. Learns from you, explains itself, works with coding agents. Successor to The Fuck.

ArcticWinterSturm / opencode-compat-shim

OpenCode Free Tier Fix for Hermes Agent

shaheeryasirofficial / ClickMyPayload

ClickMyPayload creates realistic browser verification pages that silently copy PowerShell or CMD commands to the clipboard, display decoy tokens, and guide users to execute them via Win+R. Optional binary payloads can download automatically on page load

bpm500 / zapret_pp

Zapret++ (GUI for Zapret)

cg689 / codex-reconnect-fix

一键诊断并修复 Codex / ChatGPT 桌面版「正在重新连接 1/5 → 5/5」| The Codex Rust backend ignores the Windows system proxy unless respect_system_proxy = true