An open-source macOS desktop app for DeepSeek Harness (dsh), called dsh-desktop, wraps the harness's Web UI in a native window, enabling local chat, file operations, terminal use, and approvals withou…
A developer created a skill that generates or updates an OKF-compliant index.md file for a specified directory, designed to help AI agents understand workspace structure. The skill scans only immediat…
LongHorizon-Harness, an open-source loop engineering system for AI agents, reached #1 on the Hugging Face Daily Papers weekly ranking for 2026-W32 and released v0.1.6 on 2026-08-15, adding OpenCode CL…
Cosyncing, a new open-source tool, lets users synchronize and control coding agents such as DeepSeek Harness and Claude Code from a phone or desktop via a local broker, with no hosted service. The bro…
A new open-source benchmark, holdline, measures the effectiveness of AI-agent write-guards, reporting a class-balanced Cohen's kappa of 0.82 for agreement with a 4-model judge panel on real agent traj…
Rysh, an open-source agentic terminal multiplexer written in Go, lets Claude and Codex agents work as a team in tabs, panes, and splits, with each pane acting as an agent that can answer prompts and c…
Developer Akshay Muchaklavya built Forge, a custom deep learning framework written from scratch in C++, and used it to implement GPT-2 that matches Hugging Face's transformers token-for-token under gr…
Suki, a new local-first AI tutor that integrates with coding agents, turns AI into a learning partner that forces users to restate knowledge in their own words, probe for gaps, and schedule spaced rep…
A new open-source framework called Agent Control Plane (ACP) moves AI agent authorization outside the model, into a structured-input control plane where prompt injection cannot reach, ensuring a compr…
Engelbart, a free open-source Claude Code plugin for managing goals and TODOs, is now available in early beta for macOS or Linux with Node 18+ and Claude Code 2.1.175+. The tool, created by a team foc…
Doberman, a new AI security tool launched on Hacker News, claims to stop AI models like Claude from executing dangerous commands such as deleting a database, addressing the gap left by existing guardr…
DeepSeek V4-Pro and V4-Flash models exhibit strong knowledge, reasoning, code, and agent capabilities, but their performance in complex tasks suffers from capability-realization loss due to interface …
OpenKnowledge founder Nick released Visimer, an open-source visual editor for Mermaid diagrams, enabling point-and-click editing of diagram elements while preserving file formatting. The library maps …
VectorPrism, an open-source retrieval engine from insightitsGit, claims to cut multi-vector storage costs to a single 1024-dimensional vector per chunk by multiplexing six specialized relevance subspa…
Lawrence Architectures released UL-SMF, an open-source KV-cache compression fabric that reduces memory footprint by up to 384x with over 94% semantic retention, addressing the memory bottleneck in lon…
PageLM, an open-source AI education platform developed by caviraOSS, converts study materials into interactive resources such as quizzes, flashcards, structured notes, and podcasts. Built with React, …
Foreman, an Eve software factory from Show HN, deploys AI agents across four development stations—Classifier, Analyst, Implementer, and Reviewer—to triage GitHub and Linear tasks and deliver reviewed …
A developer has created a new LLM skill called 'pragmatic-context' that enables AI models to infer sarcasm, jokes, hyperbole, roleplay, and indirect intent from conversational context. The skill instr…
Yetone released Cumora, a cross-platform team chat application where AI agents participate as first-class members alongside humans, featuring cloud-based agents powered by OpenAI's Responses API and a…
PRO-LONG, a minimal memory addition for LLM agents that appends every observation, action, and outcome to a structured log.txt and retrieves it programmatically, improves performance on the ARC-AGI-3 …