HN manuelnd 2/17/2026

How to Red Team Your AI Agent in 48 Hours – A Practical Methodology

Framework for AI red team assessments in 48 hours. Methodology for testing production AI agents with tool access against natural language attacks.

HN jeffchoi 2/17/2026

Show HN: MCP Storage Map – One MCP Server for MySQL, MongoDB, and Athena

MCP server enabling AI assistants to query MySQL, MongoDB, and AWS Athena through unified interface. Read-only by default.

HN thunderbong 2/17/2026

The Creator of OpenCode Thinks You're Fooling Yourself About AI Productivity

Interview with OpenCode creator on AI coding agents. Discusses limitations and realistic productivity expectations versus hype.

HN axjns 2/17/2026

Show HN: Llmfit;94 models, 30 providers.1 tool to see what runs on your hardware

Terminal tool that detects hardware specs and recommends which of 94 LLM models across 30 providers will run optimally.

HN ritwiktiwari 2/17/2026

Fast python project template with uv, ruff, ty and more

Opinionated Python project template with uv, ruff, pytest, MkDocs, Docker, and GitHub Actions pre-configured.

HN flatmax 2/17/2026

Show HN: ACDC – A non-agentic AI coding tool with L0-L3 context cache tiering

Terminal and browser-based AI pair-programming tool with intelligent prompt caching, context management, and structured file editing.

HN cana2026 2/17/2026

Agent Skills Hub – Security first directory for AI agent skills and MCP

Directory of 458+ security-scanned MCP skills and integrations for AI agents. Ranked by adoption and security audit scores.

HN birdculture 2/17/2026

AI is slowly munching away my passion

Personal reflection on how AI/LLM adoption is affecting programming passion and the developer community.

HN velmu 2/17/2026

Lessons learned from rebuilding a 19-year-old platform in one week with Claude

Technical case study: rebuilding 19-year-old multi-tenant platform from scratch in one week using Claude Code AI agent.

HN Zakaria_Gamal 2/17/2026

Local AI Assistant for Linux That Runs Commands

zkzkAgent: privacy-focused local AI assistant for Linux built on LangGraph and Ollama with voice interaction and command automation.

HN albertnahas 2/17/2026

Share your core values with Claude Codd every time

Claude Code plugin that injects configurable development standards into every coding session using global instructions.

HN KnuthIsGod 2/17/2026

Why AI writing is so generic, boring, and dangerous: Semantic ablation

Technical opinion on semantic ablation in LLMs: erosion of high-entropy data through greedy decoding and RLHF causing generic outputs.

HN nickjazz1998 2/17/2026

Product Lock – A JSON spec that defines what your AI-built product is and isn't

product.lock.json: JSON spec defining software product boundaries and AI management protocols. Encodes scope, review processes, decision tracking.

HN kristjansson 2/17/2026

Automated exploration of execution paths in LLM-backed applications

EnCompass: framework for AI agents separating workflow logic from search strategy. Developed with MIT researchers for exploring execution paths efficiently.

HN ronakgh97 2/17/2026

Show HN: My attempt to make Vector search engine in Rust(350k Items, ~3.5ms Qs)

Blaze-DB: High-performance vector database in Rust using HNSW indexing and SIMD. Standalone, self-managed embeddings storage.

HN WeberG619 2/17/2026

Show HN: Agent Forge – Persistent memory and desktop automation for Claude Code

Production agent framework for Claude Code with persistent memory, 17 sub-agents, desktop automation, and safety hooks.

HN linguae 2/17/2026

Fujitsu AI-Driven Software Development Platform

Fujitsu AI-Driven Software Development Platform using Takane LLM. Automates requirements, design, implementation, and testing.

HN keyhanimo 2/17/2026

AI Agents and the Rise of Judgment-Powered Software

Essay on judgment-powered AI agents autonomously taking actions. Example: agent calling owner via Twilio.

HN Jashwanth01 2/17/2026

Smart-KNN: A production-focused, feature-weighted KNN optimized for CPU

SmartKNN: production-focused KNN system with feature weighting and latency optimization for CPU deployment.

HN coolwulf 2/17/2026

Show HN: CoolWulf AI – A personal AI assistant built in Go, optimized for macOS

CoolWulf AI: self-hosted personal AI assistant in Go, optimized for macOS. Alternative to OpenClaw.

HN kqr 2/17/2026

FutureEval: Measuring the Forecasting Accuracy of AI

FutureEval: benchmark measuring AI agent forecasting accuracy across science, tech, health, geopolitics. Daily updates.

HN helloplanets 2/17/2026

Unity says its AI tech will be able to prompt full casual games into existence

Unity announces AI for game generation from prompts. Marketing announcement, limited technical depth.

Ax Joyjit Roy, Samaresh Kumar Singh 2/17/2026

Agentic AI for Commercial Insurance Underwriting with Adversarial Self-Critique

Research on agentic AI for insurance underwriting with adversarial self-critique mechanisms to ensure reliability in regulated, high-stakes environments.

Ax Lingfeng Li, Yunlong Lu, Yuefei Zhang, Jingyu Yao, Yixin Zhu, KeYuan Cheng, Yongyi Wang, Qirui Zheng, Xionghui Yang, Wenxin Li 2/17/2026

BotzoneBench: Scalable LLM Evaluation via Graded AI Anchors

BotzoneBench framework for evaluating LLMs in interactive strategic decision-making environments using graded AI anchors instead of LLM-vs-LLM tournaments.

Ax Zerui Cheng, Jiashuo Liu, Chunjie Wu, Jianzhu Yao, Pramod Viswanath, Ge Zhang, Wenhao Huang 2/17/2026

VeRA: Verified Reasoning Data Augmentation at Scale

VeRA framework for verified reasoning data augmentation at scale, converting benchmark problems into executable specifications for robust AI evaluation.

Ax Bowen Liu, Zhi Wu, Runquan Xie, Zhanhui Kang, Jia Li 2/17/2026

Scaling the Scaling Logic: Agentic Meta-Synthesis of Logic Reasoning

Research on agentic meta-synthesis of logic reasoning (SSLogic) for scaling verifiable training signals in reinforcement learning.

Ax Javier Mar\'in 2/17/2026

A Geometric Taxonomy of Hallucinations in LLMs

Taxonomy categorizing LLM hallucinations into three geometric types: unfaithfulness, confabulation, and factual error with distinct embedding signatures.

Ax Xuecong Li, Xiaohong Li, Qiang Hu, Yao Zhang, Junjie Wang 2/17/2026

Variation is the Key: A Variation-Based Framework for LLM-Generated Text Detection

VaryBalance method detects LLM-generated text by analyzing variation patterns in generated vs human text without white-box access.

Ax Truong Xuan Khanh, Truong Quynh Hoa 2/17/2026

Intelligence as Trajectory-Dominant Pareto Optimization

Theoretical framework proposing intelligence as multi-objective trajectory optimization to explain long-horizon adaptability limitations.

Ax Mayank Ravishankara 2/17/2026

PlotChain: Deterministic Checkpointed Evaluation of Multimodal LLMs on Engineering Plot Reading

PlotChain deterministic benchmark evaluates multimodal LLMs on engineering plot reading with 450 rendered plots across 15 families.

Ax Mingyang Liao, Yichen Wan, shuchen wu, Chenxi Miao, Xin Shen, Weikang Li, Yang Li, Deguo Xia, Jizhou Huang 2/17/2026

Stay in Character, Stay Safe: Dual-Cycle Adversarial Self-Evolution for Safety Role-Playing Agents

Dual-cycle adversarial self-evolution framework improves safety of LLM role-playing agents against jailbreaks while preserving persona adherence.

Ax Yuqi Xiong, Chunyi Peng, Zhipeng Xu, Zhenghao Liu, Zulong Chen, Yukun Yan, Shuo Wang, Yu Gu, Ge Yu 2/17/2026

Lang2Act: Fine-Grained Visual Reasoning through Self-Emergent Linguistic Toolchains

Lang2Act enhances Vision-Language Models with self-emergent linguistic toolchains for fine-grained visual reasoning without rigid tool separation.

Ax Rizky Ramadhana Putra, Raihan Sultan Pasha Basuki, Yutong Cheng, Peng Gao 2/17/2026

NL2LOGIC: AST-Guided Translation of Natural Language into First-Order Logic with Large Language Models

NL2LOGIC translates natural language to first-order logic using LLMs for automated reasoning in law and governance domains.

Ax Roham Koohestani, Ali Al-Kaswan, Jonathan Katzy, Maliheh Izadi 2/17/2026

AST-PAC: AST-guided Membership Inference for Code

AST-PAC applies AST-guided membership inference attacks to detect unauthorized code usage in Code LLMs trained on restricted datasets.

Ax Najmul Hasan, Prashanth BusiReddyGari 2/17/2026

DPBench: Large Language Models Struggle with Simultaneous Coordination

DPBench benchmark evaluates LLM coordination in multi-agent systems under resource contention using Dining Philosophers problem across 8 conditions.

Ax Deepak Babu Piskala 2/17/2026

MAPLE: A Sub-Agent Architecture for Memory, Learning, and Personalization in Agentic AI Systems

MAPLE proposes separating memory, learning, and personalization as distinct mechanisms in LLM agents operating on different timescales to improve user adaptation.

Ax Darren Li, Meiqi Chen, Chenze Shao, Fandong Meng, Jie Zhou 2/17/2026

General learned delegation by clones

SELFCEST: Agentic RL training enabling base models to spawn parallel clones for efficient test-time compute allocation.

Ax Muyan Weng, Defu Cao, Wei Yang, Yashaswi Sharma, Yan Liu 2/17/2026

TemporalBench: A Benchmark for Evaluating LLM-Based Agents on Contextual and Event-Informed Time Series Tasks

Benchmark evaluating LLM agents on temporal reasoning and time series tasks with progressive contextual and event information.

Ax Rohan Subramanian Thomas, Shikhar Shiromani, Abdullah Chaudhry, Ruizhe Li, Vasu Sharma, Kevin Zhu, Sunishchal Dev 2/17/2026

ProMoral-Bench: Evaluating Prompting Strategies for Moral Reasoning and Safety in LLMs

Unified benchmark evaluating 11 prompting paradigms across LLM families for moral reasoning and safety alignment.

Ax William Waites 2/17/2026

Artificial Organisations

Proposes organizational/institutional design patterns for multi-agent AI systems to achieve reliability through compartmentalization and adversarial review.

Ax Hanchen David Wang, Clayton Cohn, Zifan Xu, Siyuan Guo, Gautam Biswas, Meiyi Ma 2/17/2026

BEAGLE: Behavior-Enforced Agent for Grounded Learner Emulation

LLM-based agent simulating authentic student learning behaviors in open-ended problem-solving for education research.

Ax Yifan Ding, Yuhui Shi, Zhiyan Li, Zilong Wang, Yifeng Gao, Yajun Yang, Mengjie Yang, Yixiu Liang, Xipeng Qiu, Xuanjing Huang, Xingjun Ma, Yu-Gang Jiang, Guoyu Wang 2/17/2026

Mirror: A Multi-Agent System for AI-Assisted Ethics Review

Multi-agent system using LLMs to assist with ethics review and research governance decision-making.

Ax Daesik Jang, Morgan Lindsay Heisler, Linzi Xing, Yifei Li, Edward Wang, Ying Xiong, Yong Zhang, Zhenan Fan 2/17/2026

DECKBench: Benchmarking Multi-Agent Frameworks for Academic Slide Generation and Editing

Benchmark for evaluating multi-agent frameworks on academic slide generation and iterative editing tasks.

Ax Jisung Shin, Daniel Platnick, Marjan Alirezaie, Hossein Rahnama 2/17/2026

Situation Graph Prediction: Structured Perspective Inference for User Modeling

Frames user perspective modeling as inverse inference problem using situation graphs and ontology-aligned representations.

Ax Flint Xiaofeng Fan, Cheston Tan, Roger Wattenhofer, Yew-Soon Ong 2/17/2026

Information Fidelity in Tool-Using LLM Agents: A Martingale Analysis of the Model Context Protocol

Theoretical framework analyzing error accumulation in Model Context Protocol agents, proving linear growth in cumulative distortion.

Ax Tri Nguyen, Huy Hoang Bao Le, Lohith Srikanth Pentapalli, Laurah Turner, Kelly Cohen 2/17/2026

Detecting Jailbreak Attempts in Clinical Training LLMs Through Automated Linguistic Feature Extraction

Detects jailbreak attempts in clinical LLMs using automated linguistic feature extraction instead of manual annotation.

Ax Michael Winikoff 2/17/2026

Contrastive explanations of BDI agents

Extends BDI agent explanation mechanism to support contrastive questions about agent decision-making choices.

Ax Chen Yang, Guangyue Peng, Jiaying Zhu, Ran Le, Ruixiang Feng, Tao Zhang, Xiyun Xu, Yang Song, Yiming Jia, Yuntao Wen, Yunzhi Xu, Zekai Wang, Zhenwei An, Zhicong Sun, Zongchao Chen 2/17/2026

Nanbeige4.1-3B: A Small General Model that Reasons, Aligns, and Acts

Open-source 3B parameter language model achieving agentic behavior, code generation, and reasoning through reward modeling.

Ax Simon Rosen, Siddarth Singh, Ebenezer Gelo, Helen Sarah Robertson, Ibrahim Suder, Victoria Williams, Benjamin Rosman, Geraud Nangue Tasse, Steven James 2/17/2026

MoralityGym: A Benchmark for Evaluating Hierarchical Moral Alignment in Sequential Decision-Making Agents

Benchmark for evaluating moral alignment in sequential decision-making agents using hierarchical deontic constraints and trolley-dilemma scenarios.

Ax Anhao Zhao, Ziyang Chen, Junlong Tong, Yingqi Fan, Fanghua Ye, Shuhao Li, Yunpu Ma, Wenjie Li, Xiaoyu Shen 2/17/2026

On-Policy Supervised Fine-Tuning for Efficient Reasoning

Proposes on-policy supervised fine-tuning as simpler alternative to RL for training reasoning models with better efficiency.