Proof-Carrying Messages: a protocol for AI agents whose claims carry evidence the receiver checks offline

TestOPS一站式AI运维平台 · By 一鸣（公众号：一鸣AI编程）

Benchmark aberto de LLMs para o processo legislativo brasileiro

端到端智能体平台：把本机的 AI Agent 变成人人可用的 Web 产品（Vue 3 + Flask + Go WS 网关 + Docker 部署）

A decision library for running local LLMs on multi-GPU machines without NVLink: parallelism, communication, quantization, speculative decoding, KV cache. Every conclusion carries its evidence grade and reproducible conditions.

The CAD runtime built for AI agents: 2D drawings (DXF/DWG) and parametric 3D parts (FreeCAD, STEP) through MCP, CLI and Python, with undo, independent validation and a live viewer.

Unofficial read-only MCP server for Iranian job sites (استخدام) IranTalent, Jobinja and Jobvision: job search, companies, reviews, interview reports and pay side by side. For Claude, Cursor, Copilot and AI agents.

Research internship cold-email tool: finds professors matching your resume, writes personalised emails, sends them paced from your inbox, tracks replies. Free, open source. By Avinash Chandra Gupta.

Policy-adherent, auditable reasoning for AI Agents.

网文平台资源配置决策助手：GraphRAG 两阶段检索 + ReAct Agent + 规则引擎兜底，方案合法率 60%→100%

Topology optimization for 3D printing, as an MCP server for Claude, Codex, Cursor and other AI agents. Loads and regions in, a lighter printable STL out.

Local personal-data firewall for cloud LLMs. Scrubs names, emails, IDs and secrets before they leave your machine.

ComfyUI nodes for Strata: text, structured prompts, vision and managed local inference with GPU handoff. Publisher: t8star.

Move a personal AI agent between platforms without losing its job, memory, or schedules. A neutral dossier format with adapters for Grok Bot, Muse, Hermes, and OpenClaw.

Benchmark and fix LLM hallucinations when extracting figures from SEC 10-K filings

A state machine protocol for running controlled ML experiments with coding agents.

A reference chart of neural network architectures, from the perceptron to diffusion LLMs, drawn from one spec as a poster and an interactive page. Inspired by the Neural Network Zoo.

Self-hosted Google Analytics alternative that reads your server access log: real people, AI assistants, crawlers, bots and the scanners probing your site, which JavaScript analytics never see. Leads and revenue by source. Every figure counted twice. No cookies, no third-party script. Python, SQLite, Docker.

A skill for AI coding Agents; Free OEM service manuals -> procedure, torque table, labelled figures. 69 makes, 1960-2025. Agent skill + stdlib navigator.

An experiment-driven LLM engineering lab with 9 reproducible Python experiments exploring token economics, API costs, caching, context growth, latency, model routing, rate limits, concurrency, and batching using Groq.

Bulk classification and decision inference on one GPU: the content is read once, every question about it is scored in the same pass. OpenAI-compatible, Apache 2.0

Parametric CAD kernel for LLMs: spec-driven geometry, measured facts, three-state verification, Blender-backed delivery