Navox Agents provides 7 specialist AI agents for Claude Code with human-in-the-loop checkpoints. Demonstrated building Rust project and game with debugging and testing without external platform.
RepoGauge is an open-source tool for comparing AI coding agent performance on real repositories, addressing token costs and model efficiency variations across providers.
Smith orchestrates multiple AI coding agents (Claude, Codex, Gemini, Aider) in parallel with git worktrees, live streaming output, and MCP server integration for shared team configuration.
Code AI agent system supporting auto, asynchronous, concurrent execution with local model support and 100% local data processing. Limited documentation.
TileTensor: Mojo's tensor type for expressing complex GPU memory layouts safely and efficiently in high-performance kernel development.
AAIP protocol proposal for verified AI agent identity and autonomous agent-to-agent commerce, enabling agents to operate on web without human impersonation.
Technical writeup on security practices for agentic AI using sandboxes and worktrees in 2026. Title-only stub.
Dome Systems framework for controlling and managing AI agent behavior. Title-only stub.
Article on Amazon's emphasis on Model Context Protocol (MCP) as agentic AI adoption accelerates. Title-only stub.
AgentFM is peer-to-peer network enabling distributed AI workload processing across idle CPUs/GPUs without centralized cloud providers.
GitHub CLI tool (gh skill) for discovering, installing, and publishing agent skills across multiple platforms including Copilot, Claude Code, Cursor.
MCP server converting documentation URLs into Claude skill files for AI agents. No API key required. Supports Claude Code, Cursor, Windsurf.
Claude Code plugin detecting and fixing 14 types of agent quality issues via trace analysis. Improves agent reliability and context handling.
Request for feedback on paper about revision-capable language models. Minimal content provided.
Web analytics API for AI agents to query performance data, detect regressions, and surface insights via chat. Open source, Next.js/React compatible.
Benchmark for evaluating PDF document parsing tools that AI agents depend on. Tests ~2K real enterprise documents across failure modes.
Analysis of Claude Opus fixing production bugs with technically correct but wrong solutions. Explores model capability gaps between diagnosis and root cause analysis.
MacBook notch widget displaying Claude Code dashboard metrics and status. UI utility for monitoring Claude usage.
Tracker showing which developer tools and libraries each LLM model selects across thousands of prompts and skill levels.
Open source job aggregator app with MCP apply-agent server for automated job applications. React/Node/SQLite stack with agent workflow support.
Analysis of Claude Opus 4.7 release and developer sentiment shift. Author discusses coding AI tool reliability and trust issues.
Essay on AI agents in education settings. Discusses adoption and student experience without technical depth.
Mimikos: Zero-config mock server that generates deterministic API responses from OpenAPI specs, useful for testing.
Method to read Claude Code quota from local config files without API calls.
CLIver: Terminal AI agent with customizable prompts, skills, MCP integrations, and permission controls for coding, DevOps, and analysis tasks.
Kubernetes operator for deploying and managing self-hosted LLM inference engines with Ollama and OpenAI-compatible APIs.
One-week hackathon challenge to build projects using resilient LLM framework. Limited details on scope or prizes.
Video about a statistical paradox relevant to machine learning. No details on content or originality.
Cross-tool AI memory infrastructure enabling context persistence across Claude, ChatGPT, Cursor and other coding assistants.
Open-source (MIT) shared memory and coordination layer for multiple Claude Code instances. Runs locally, Node.js setup.
Solo developer built full-stack charitable giving SaaS using agentic AI to accelerate development and reduce complexity.
Discussion of context limitations in LLM agents, specifically Cursor IDE's agent behavior changes and developer frustration.
Discussion of authorization challenges in AI agent deployments, beyond identity management issues.
Tool to audit website readiness for AI agents, checking robots.txt, MCP, OAuth, and agentic commerce standards.
Lazyagent is a terminal TUI that collects and visualizes events from Claude Code and other AI runtimes, helping developers track what agents are doing and spot when execution goes off track.
Show HN: AI agent system that refines ML models for Zephyr RTOS embedded systems. Demonstrates agent-driven ML optimization.
Claude MCP server converting saved articles into AI-generated two-host podcast conversations stored locally as MP3.
WorldSeed: Open-source world engine for defining scenarios in YAML and running AI agents autonomously within them.
Enterprise security risks from autonomous AI agents exposing data at machine speed; traditional controls inadequate.
Guide on inference caching techniques in LLMs for improved performance. Title-only, limited detail available.
Offline CLI tool that analyzes Spring Boot and Java logs to identify errors, root causes, and fixes without dependencies.
Research on LLM compression achieving 22% size reduction without quality loss. Model optimization technique.
AI tool converting audio/YouTube links to playable piano sheet music in PDF, MusicXML, and MIDI formats.
Hands-on tutorial fine-tuning Qwen3-4B LLM with LoRA on M1 Mac for customer support, including synthetic data generation and local inference.
Open-source protocol for tracking AI agent task commitments and proof of delivery.
Analysis of agentic AI ecosystem developments and Java 26 language features for agents.
Gradient-free adaptation of 1-bit language models via discrete search with binary weight groups and targeted patching.
Adapt: LLM-based adaptive memory layer for AI apps that restructures itself based on data patterns. Self-evolving storage alternative to RAG.
Mabon.ai: AI agent continuously searches job boards with intent understanding for targeted matches.
Dashbit explores code review workflow improvements for human-agent collaboration with coding agents.