3/5/2026
Show HN: Sokuji – Open-source speech translator with on-device AI WASM/WebGPU
Open-source speech translation app with on-device ASR/TTS using WASM/WebGPU, no internet required.
3/5/2026
Show HN: Entropy – API security testing using LLMs to generate attack scenarios
Entropy uses LLMs to analyze API schemas and generate custom adversarial attack sequences for security testing beyond static scanners.
3/5/2026
Comprehensive check and Productivity tool for vibecoders. Has a team mode too
Developer tool for verifying AI-generated code diffs across Claude Code, Cursor, and other AI coding assistants. Includes team mode and CI integration.
3/5/2026
Show HN: Selva – Amazon for AI Agents
Selva is a skill for AI agents to search, compare, and purchase products on Amazon, with secure Stripe integration and spending limits.
3/5/2026
Show HN: Montage – Quickly build product launch videos with coding agents
Montage: Remotion fork enabling coding agents to rapidly generate product launch videos by automating motion graphics using reusable animation primitives.
3/5/2026
Show HN: OXPT – Visual branching canvas for prompt versioning (Korean support)
OXPT is a visual branching canvas for prompt versioning and non-linear exploration, supporting complex patterns like self-reflection and tree-of-thought with Korean support.
3/5/2026
A 130KB Markdown file that turns Claude Code into an opinionated senior PM
130KB Markdown knowledge base that configures Claude Code as an opinionated PM agent with 6 domains, 30+ frameworks, and 12 templates for product strategy.
3/5/2026
Show HN: Beads planner plugin for Claude Code
Beads planner plugin for Claude Code decouples planning from execution, integrating Steve Yegge's Beads methodology with Claude's planning mode.
3/5/2026
Datasets for Reconstructing Visual Perception from Brain Data
Open repository indexing neuroimaging fMRI datasets for reconstructing visual perception from brain data, with guidance for AI/ML researchers new to neuroscience methods.
3/5/2026
Show HN: Claude Code plugin that adds CRDT collaboration to any app in 10 min [video]
Claude Code plugin for Velt that installs real-time CRDT collaboration features (sync, comments, presence) into apps in under 10 minutes.
3/5/2026
Preventing MCP rug pull attacks that steal AI agent credentials
Security measures to prevent MCP rug pull attacks that compromise AI agent credentials.
3/5/2026
The 89% Problem: LLMs are resurrecting the "dormant majority" of open source
Analysis of how LLMs are reviving dormant open source packages, creating security risks as 89% of downloads may use abandoned code.
3/5/2026
Show HN: Todoglow – Keyboard-first todo app for macOS with MCP support
Todoglow: keyboard-first macOS todo app with MCP support that tracks atomic task durations and integrates with task trackers like Notion and Jira.
3/5/2026
AI Twitter's favourite lie: everyone wants to be a developer
Critique arguing AI-powered coding doesn't make everyone a developer; most users lack software-building motivation.
3/5/2026
Show HN: Attn – Markdown viewer and editor in a <20MB binary (Rust)
Attn: lightweight Markdown viewer/editor built in Rust (<20MB) designed to work with Claude Code for reading planning docs and architecture notes from terminal.
3/5/2026
Show HN: LiberClaw, deploy AI agents that run 24/7 on their own VMs
Open-source platform deploying AI agents on dedicated VMs with markdown-defined skills and 24/7 execution.
3/5/2026
Show HN: OmoiOS–190K lines of Python to stop babysitting AI agents (Apache 2.0)
190K-line Python framework for managing AI coding agents with monitoring, drift detection, and task validation.
3/5/2026
Show HN: ConvertlyAI – Format raw transcripts into 10 social assets
LLM-powered tool converting raw transcripts into formatted social media content (Twitter, LinkedIn, etc).
3/5/2026
Show HN: ExtraSuite – pull / push for Google Drive
ExtraSuite enables AI agents to make targeted edits to Google Drive files (sheets/docs/slides) using a terraform-like workflow approach.
3/5/2026
Show HN: Reduce LLM token use by ~30% with this MCP/CLI tool(Claude benchmarked)
CLI tool using tree-sitter and ripgrep to reduce LLM token usage by ~30% via smart code context selection.
3/5/2026
HATEOAS Works with an LLM in the Mix
Discussion of HATEOAS REST API design patterns and how LLMs could enable smarter API clients for adaptive discovery.
3/5/2026
Show HN: Webmcp-react – React hooks that turn your website into an MCP server
React hooks library enabling websites to expose typed tools to AI agents via navigator.modelContext browser API.
3/5/2026
Show HN: Pre-execution verification for LLM-generated agentic workflows
workflow-verify adds a verification layer for LLM-generated agentic workflows before production execution, validating types and side effects to prevent data corruption.
3/5/2026
Defensive Refusal Bias in LLMs Is Hurting Infosec
Analysis of how defensive refusal mechanisms in LLMs negatively impact infosecurity capabilities.
3/5/2026
Show HN: OpenRouter Skill – Reusable integration for AI agents using OpenRouter
Open-source reusable skill for AI agents integrating OpenRouter with model discovery, routing, and multimodal chat patterns.
3/5/2026
Flyte 2 In-Browser Demo: Open-Source AI Orchestration Is Now Available Locally
Flyte 2 releases in-browser demo for open-source ML orchestration platform with local execution.
3/5/2026
The ML Engineer's Guide to Protein AI
Guide for ML engineers on protein AI applications and techniques.
3/5/2026
Show HN: SamarthyaBot – a privacy-first self-hosted AI agent OS
SamarthyaBot: Self-hosted privacy-first AI agent OS with terminal, SSH, browser control, and multi-LLM support.
3/5/2026
AI Agent Authentication and Authorization IETF RFC Draft
IETF RFC draft for AI agent authentication and authorization standards.
3/5/2026
In Which We Give Our AI Agent a Map (and It Stops Getting Lost)
Research on improving AI agent navigation and planning by providing environmental map context.
3/5/2026
Doing My Taxes with Claude
Personal evaluation of Claude's capabilities on tax preparation tasks, demonstrating real-world LLM application.
3/5/2026
Show HN: Cook – A portable terminal AI agent (OSS, MIT)
Open-source portable terminal AI agent with MIT license for command-line task execution.
3/5/2026
Show HN: ClawLite – AI agent runtime built on Android Termux
AI agent runtime built on Android Termux platform.
3/5/2026
Show HN: Neo – AI-powered native .NET desktop app generator
AI-powered tool generating compiled .NET desktop applications from natural language prompts without mockups.
3/5/2026
How Easy Is It to Trick an AI? Notes from a Red Team Competition
Writeup from red team competition testing adversarial attacks against AI systems.
3/5/2026
Show HN: Merkle Mountain Range audit log and execution tickets for AI agents
Cryptographic runtime integrity system using Merkle Mountain Range to create tamper-evident audit logs for agentic AI systems.
3/5/2026
Show HN: Rivora, Reliability for Autonomous Infrastructure
Reliability layer for autonomous infrastructure with proactive failure detection and AI-aware deployment tracking.
3/5/2026
Show HN: DevOpsAgents – AI agents to deploy and manage your infra
DevOps Agents: specialized AI agents that analyze GitHub repos, deploy cloud resources, and manage infrastructure via chat interface.
3/5/2026
Show HN: Meto – Methodology backbone for AI agentic coding
CLI tool scaffolding structured multi-agent projects with kanban boards and coding conventions for organized Claude Code sessions.
3/5/2026
AI Is Confidently Wrong
Benchmark study showing 72 AI models' failure rates on nonsense detection, with ChatGPT pushing back 27% vs Gemini 10%.
3/5/2026
Chasing the AI High: Clay, Kilns, and the Red Queen's Race
Opinion piece on AI agents in software development and generative transformers, lacking specific technical details or data.
3/5/2026
I miss the grind of writing software before AI
Personal essay on how LLM coding assistants reduce learning through faster task completion, reflecting on trade-offs between speed and mastery.
3/5/2026
Show HN: RAGLight, serve a RAG pipeline as a REST API and chat UI in one command
Open source RAG library with new serve command to deploy modular RAG pipeline as FastAPI REST API and Streamlit chat UI with single command.
3/5/2026
Guardrailing intuition: Towards reliable AI
Exploration of symbolic AI vs connectionist approaches and their implications for reliable AI systems.
3/5/2026
Why AI Agent Workflows Need Governance
Stub article on governance requirements for AI agent workflows.
3/5/2026
Show HN: MDMA – Markdown extension that makes LLM responses interactive
MDMA: Markdown extension enabling LLMs to generate interactive components (forms, tables, webhooks) instead of plain text responses.
3/5/2026
Show HN: Sanna – OpenClaw for your phone. Open-source voice AI agent for Android
Sanna: open-source voice AI agent for Android enabling task management and device control hands-free, alternative to Siri/Google Assistant.
3/5/2026
What changes in engineering teams once AI tools "click"?
Case study on how engineering teams adopt AI tools in real workflows: skepticism, gradual trust-building, and productivity changes over 2 days.
3/5/2026
Show HN: Stop LLMs from brute forcing (guessing) APIs
Developer addresses problem of AI agents brute-forcing and guessing API interfaces despite complete documentation, proposing client-side fine-tuning solutions.
3/5/2026
Show HN: The AI agent that grows your YouTube channel on autopilot
CLI and AI agent for YouTube automation including video uploads, optimization, comment management, and channel branding using MCP server.