3/6/2026
Show HN: Evalcraft – cassette-based testing for AI agents (pytest, $0/run)
Pytest plugin using cassette-based recording for AI agent testing. Records LLM calls once, replays deterministically in CI with zero API costs.
3/6/2026
OpenCAD: Modular CAD System
Modular parametric CAD system with AI-assisted design capabilities built with Python and Node.js.
3/6/2026
World Monitor – AI-powered news aggregation
Real-time global intelligence dashboard with AI-powered news aggregation, geopolitical monitoring, and multi-model summarization chain.
3/6/2026
AgentSeal – Security scanner for AI agents (150 attack probes)
Security scanner testing AI agents against 150+ attack probes including prompt injection and tool misuse vulnerabilities.
3/6/2026
Mcpup – Manage MCP servers across 13 AI clients from one config
CLI tool managing Model Context Protocol servers across 13 AI clients from single config. Includes 97 curated MCP server templates.
3/6/2026
OpenClaw on Amazon Lightsail to run your autonomous private agents
OpenClaw general availability on Amazon Lightsail: open-source self-hosted autonomous AI agent with browser integration and Bedrock model provider.
3/6/2026
Show HN: I built an AI-powered advert – is this the future of advertising?
Interactive AI-powered sales chat demo for fictional luxury EV brand. Explores conversational advertising with product knowledge retrieval.
3/6/2026
Coding with Multiple AI Agents to Build Scalable Rate-Limiting Infrastructure
Case study: Ayrshare's dev team used multi-agent Claude work via Cursor to build scalable rate-limiting infrastructure for social media APIs.
3/6/2026
How Descript enables multilingual video dubbing at scale
Descript uses OpenAI models for multilingual video dubbing, optimizing translations for meaning and timing.
3/6/2026
Show HN: NeoMud – A multiplayer dungeon game with AI agents that QA and playtest
MUD game built with Claude Code featuring agent pipeline: game-designer agent proposes balance changes, playtest agent QA-tests gameplay.
3/6/2026
Ask HN: Are subagents making flagship AI models feel like managers?
Discussion questioning whether subagents in AI coding assistants create manager-like behavior, reducing flagship models' reasoning autonomy.
3/6/2026
Show HN: ParseForce – I got tired of brittle email parsers so I built this
LLM-based email parser tool handling brittle format variations with schema validation, retries, and backend integration.
3/6/2026
Show HN: Introducing Kite AI Agent: Conversational Operations for Kubernetes
Conversational AI agent for Kubernetes cluster operations. Combines visual dashboard with agentic workflows to reduce context switching between monitoring and CLI tasks.
3/6/2026
Online harassment is entering its AI era
Open-source maintainers face harassment from AI-generated code contributions. matplotlib and other projects implement human review policies for AI-submitted code.
3/6/2026
Show HN: Claude Code for iPad – Agentic AI coding tool with file ops, Git, shell
iPad agentic coding tool using Claude that autonomously reads codebases, plans changes, edits files, and pushes to GitHub. 7 integrated tools execute locally with token-streaming API.
3/6/2026
Don't Be a Wrapper, Be a Container
Hopsworks platform as runtime for coding agents. Analysis of container vs wrapper approaches and MCP integration for Claude Code, Codex, Gemini.
3/6/2026
Show HN: Koshei AI – a voice-native AI language university (A1 to D2)
Voice-first AI language learning application built with Next.js and Gemini APIs. Technical discussion of browser-native lip sync implementation challenges.
3/6/2026
Show HN: Codebase-md – Creates Claude.md, .cursorrules, AGENTS.md from any repo
Tool that generates unified context files (CLAUDE.md, .cursorrules, etc.) for AI coding tools from a single codebase scan, solving maintenance overhead.
3/6/2026
Kuberna Labs: AI's Economic Engine
Platform for building and deploying autonomous AI agents across decentralized networks. Combines learning environment with infrastructure for cross-chain finance and trusted execution.
3/6/2026
Shellfirm – Safety guardrails for AI coding agents
Safety guardrails tool intercepting dangerous commands from both humans and AI agents. Works universally across shells with blast radius detection and context analysis.
3/6/2026
Show HN: 1v1 coding game that LLMs struggle with
Side project: 1v1 strategy game where LLMs compete. Results show LLMs can generate functioning bots but underperform vs. human players.
3/6/2026
Humans and Agents in Software Engineering Loops
Article exploring optimal human-agent collaboration in software development. Discusses whether developers should inspect all code or focus on outcomes.
3/6/2026
Zammad open-source helpdesk introduces AI without LLM lock-in
Open-source helpdesk software version 7.0 adds AI features without vendor lock-in. Supports pluggable LLMs while maintaining data sovereignty.
3/6/2026
WavSLM: Single-Stream Speech Language Modeling via WavLM Distillation
WavSLM: single-stream speech language model using WavLM distillation for autoregressive speech generation without text supervision.
3/6/2026
GALACTIC: Global and Local Agnostic Counterfactuals for Time-series Clustering
GALACTIC: counterfactual explanations method for time-series clustering identifying transitions across cluster boundaries.
3/6/2026
FairFinGAN: Fairness-aware Synthetic Financial Data Generation
FairFinGAN: WGAN-based framework for generating synthetic financial data with fairness constraints to mitigate bias.
3/6/2026
Preserving Continuous Symmetry in Discrete Spaces: Geometric-Aware Quantization for SO(3)-Equivariant GNNs
Geometric-aware quantization technique preserving SO(3)-equivariance in GNNs for efficient molecular simulations.
3/6/2026
InfoFlow KV: Information-Flow-Aware KV Recomputation for Long Context
InfoFlow KV method for efficient long-context RAG by recomputing KV caches with information-flow awareness to optimize prefilling.
3/6/2026
Learning Causal Structure of Time Series using Best Order Score Search
TS-BOSS: time series extension of Best Order Score Search for causal structure discovery in multivariate temporal data.
3/6/2026
Embedded Inter-Subject Variability in Adversarial Learning for Inertial Sensor-Based Human Activity Recognition
Deep adversarial framework for human activity recognition from wearable sensors addressing inter-subject variability.
3/6/2026
Robust Node Affinities via Jaccard-Biased Random Walks and Rank Aggregation
TopKGraphs method for node similarity estimation using Jaccard-biased random walks for clustering and recommendation tasks.
3/6/2026
On the Necessity of Learnable Sheaf Laplacians
Sheaf Neural Networks extension with learnable restriction maps to address oversmoothing on heterophilous graphs.
3/6/2026
An interpretable prototype parts-based neural network for medical tabular data
Interpretable prototype parts-based neural network for medical tabular data with discretization of diagnostic norms for improved trust and transparency.
3/6/2026
Latent Wasserstein Adversarial Imitation Learning
LWAIL framework for imitation learning from state-only demonstrations using adversarial Wasserstein methods in latent space.
3/6/2026
Censored LLMs as a Natural Testbed for Secret Knowledge Elicitation
Study of honesty elicitation and lie detection methods in LLMs using censored models as natural testbed for secret knowledge extraction.
3/6/2026
POET-X: Memory-efficient LLM Training by Scaling Orthogonal Transformation
POET-X framework for memory-efficient LLM training using orthogonal transformations to improve stability while reducing memory consumption.
3/6/2026
One Size Does Not Fit All: Token-Wise Adaptive Compression for KV Cache
Token-wise adaptive compression method for KV cache in LLMs to reduce memory footprint during inference without expensive retraining or performance loss.
3/6/2026
Context-Dependent Affordance Computation in Vision-Language Models
Vision-language models show massive context-dependent affordance drift when primed with different agentic personas, affecting action prediction.
3/6/2026
When Scaling Fails: Network and Fabric Effects on Distributed GPU Training Performance
Empirical study showing distributed GPU training scaling fails due to network and fabric effects overlooked by training frameworks.
3/6/2026
AMV-L: Lifecycle-Managed Agent Memory for Tail-Latency Control in Long-Running LLM Systems
AMV-L: memory lifecycle management system for long-running LLM agents that bounds tail latency through intelligent retention policies.
3/6/2026
SkillNet: Create, Evaluate, and Connect AI Skills
SkillNet: open infrastructure for creating, evaluating, and sharing reusable AI skills to enable systematic knowledge accumulation across agents.
3/6/2026
Induced Numerical Instability: Hidden Costs in Multimodal Large Language Models
Multimodal LLMs fail via induced numerical instability when adversarial images trigger optimization of inference-time instability.
3/6/2026
Act-Observe-Rewrite: Multimodal Coding Agents as In-Context Policy Learners for Robot Manipulation
Act-Observe-Rewrite: multimodal LLM agent learns robot manipulation by synthesizing Python controller code between trials via visual feedback without gradient updates.
3/6/2026
AbAffinity: A Large Language Model for Predicting Antibody Binding Affinity against SARS-CoV-2
LLM-based approach for predicting antibody binding affinity against SARS-CoV-2 using machine learning on experimental antibody data.
3/6/2026
Self-Attribution Bias: When AI Monitors Go Easy on Themselves
AI agents using LLMs for self-monitoring exhibit self-attribution bias when evaluating their own actions, showing leniency compared to external evaluation.
3/6/2026
Spinverse: Differentiable Physics for Permeability-Aware Microstructure Reconstruction from Diffusion MRI
Spinverse: differentiable physics method for reconstructing tissue microstructure from diffusion MRI using Bloch-Torrey simulation.
3/6/2026
iAgentBench: Benchmarking Sensemaking Capabilities of Information-Seeking Agents on High-Traffic Topics
iAgentBench: benchmark for evaluating information-seeking AI agents on cross-source sensemaking and evidence integration tasks.
3/6/2026
Improving the accuracy of physics-informed neural networks via last-layer retraining
Method improving physics-informed neural network accuracy through last-layer retraining on PDE solutions.
3/6/2026
When Denoising Hinders: Revisiting Zero-Shot ASR with SAM-Audio and Whisper
Empirical study examining whether speech denoising improves zero-shot ASR with SAM-Audio and Whisper models.
3/6/2026
SLO-Aware Compute Resource Allocation for Prefill-Decode Disaggregated LLM Inference
Method for allocating compute resources across prefill-decode disaggregated LLM inference while meeting SLO constraints.