Safer agent skills

Reduce LLM costs by automatically selecting the most efficient model using real benchmark data and a single easy-to-use plugin.

C · Review first · Execution High

Monitor LLM token costs in real time with a terminal dashboard offering per-request tracking, budget control, and alerting without external services.

C · Review first · Execution High

Desktop chat app for Microsoft's 1-bit BitNet LLMs. Windows-native, CPU-only, zero dependencies

C · Review first · Execution High

Estimate token usage and API costs for large language models to help developers manage AI prompt expenses before scaling workloads.

C · Review first · Execution High

Provide a checklist to ensure secure, compliant, and well-governed AI and LLM integration for responsible application development.

C · Review first · Execution High

Transform probabilistic LLMs into deterministic state machines to improve logic, prevent context loss, and generate unique, reliable responses.

C · Review first · Execution High

Hosts LLM model metadata for the LLM Architecture Gallery, giving a clean public data source for model cards, dates, images, and links

C · Review first · Execution High

途正英语课前AI分级测评系统，采用AI语音实时对话模式，通过自适应出题算法（CAT）精准评定学员英语水平（零级/一级/二级/三级）。以微信小程序交互体验为标准，包含欢迎引导页、规则说明页、AI语音对话测评核心页面、测评结果展示页面。集成LLM大模型评估、TTS语音合成、ASR语音识别。

C · Review first · Execution High

Evaluate and compare LLM responses across five key quality dimensions to improve post-training feedback and analysis.

C · Review first · Execution High

Extract structured data from websites using LLM-powered semantic extraction that adapts automatically to site changes for fast, efficient scraping.

C · Review first · Execution High

Automate Chrome with any LLM provider to click, type, scroll, read pages, take screenshots, and replay browser workflows.

C · Review first · Execution High

Build runtime reliability guards for LLM pipelines with one wrap() call and configurable protection against common production failures

C · Review first · Execution High

Build a text-based alchemy game with Ollama that mixes elements into new ones using a local LLM and saves your progress

C · Review first · Execution High

Automate XJTLU email search, extract full message text, and generate Chinese AI summaries with Playwright, FastAPI, and LLM APIs

C · Review first · Execution High

Optimize Tesla V100 GPUs for AWQ 4-bit inference with improved speed, stability, and support for modern large models like Qwen3.5 and MoE.

C · Review first · Execution High

Train and chat with distributed federated LLMs using community GPU power and smart model routing

C · Review first · Execution High

Build local video captions with frame-by-frame vision inference using React, Express, and MLX on Apple Silicon Macs

C · Review first · Execution High

GitHub repository Ayushmore1214/llm-k8s-deployment; metadata recovered from GitHub Search API.

C · Review first · Execution High

Evaluate LLM safety, fairness, and reliability with an open-source framework for regulated industries

C · Review first · Execution High

Guide fine-tuning, quantization, and evaluation for LLMs, vision-language models, and embedding models from theory to practice

C · Review first · Execution High

LLM WIKI

C · Review first · Execution High

Store emotional memories and personality profiles for LLMs to deliver real-time emotion detection and tailored advice

C · Review first · Execution High

Find and duplicate transformer reasoning circuits to boost LLM logic without training or weight changes

C · Review first · Execution High