3/2/2026
Infrastructure Agents Guide – Design and operate AI agents for infra safely
Guide on designing and operating AI agents for infrastructure management. Covers spectrum from copilot-mode suggestions to autonomous agentic deployments.
3/2/2026
Pydantic Monty: A Minimal Python Sandbox for AI Agents
Pydantic Monty: open-source minimal Python sandbox in Rust for safely executing code written by AI agents. Enables efficient code generation vs tool calls.
3/2/2026
A webpage that supports controlling 16+ CLI Agents(CC,Codex,GeminiCli) remotely
Web interface for controlling 16+ CLI AI agents remotely. Limited technical detail provided.
3/2/2026
Show HN: AgentEmail Free and Open Source Email Infra for AI Agents
Open source Bun CLI providing disposable email inbox access for AI agents, MIT licensed with JSON output.
3/2/2026
Show HN: Oc-mnemoria – Persistent memory for AI coding agents
OpenCode plugin providing persistent shared memory for AI coding agents across sessions, storing decisions and discoveries in Rust backend.
3/2/2026
DystopiaBench – Measuring AI's willingness to ruin humanity
Research benchmark measuring AI model compliance with dystopian directives across 5 escalation levels in infrastructure, weapons, and surveillance domains.
3/2/2026
Show HN: LLM Evaluator for "Who is hiring" threads
CLI tool using Google Gemini to evaluate job postings from HN's Who's Hiring threads against user preferences and resume.
3/2/2026
Perspective Server
macOS app exposing Apple's on-device Foundation Models via OpenAI-compatible local API for running AI without cloud services.
3/2/2026
Show HN: Open-Jet – self-hosted Agentic TUI for air-gapped Jetsons
Terminal UI for self-hosted AI agents on Jetson edge devices, optimized for air-gapped environments with 17 tok/s on Orin Nano.
3/2/2026
Show HN: GitAgent – Clone a repo, get an AI agent – Claude Code / OpenClaw
GitAgent: framework-agnostic open standard for defining AI agents via repository configuration. Single command deploys agents across Claude, OpenAI, or custom backends.
3/2/2026
Show HN: Ragtoolina – MCP tool that adds codebase RAG to AI coding agents
Ragtoolina MCP server: pre-indexes codebases for AI coding agents. Benchmarked 63% token reduction and 43% fewer tool calls vs file scanning approach.
3/2/2026
Show HN: Claw – Let your AI agent operate any machine as if it were local
Claw: MCP server enabling AI agents remote machine access via SSH. Provides bash, file operations, and editing tools without opening ports or requiring root.
3/2/2026
RAG vs. Skill vs. MCP vs. RLM
Comparison of four techniques for extending LLM capabilities: RAG, RLM, SKILL, and MCP. Covers implementation mechanics, tradeoffs, and usage guidance.
3/2/2026
Show HN: C7 – Pipe up-to-date library docs into any LLM from the terminal
Technical comparison of RAG, Skill, MCP, and RLM approaches for extending LLM capabilities: context expansion and external tool access patterns.
3/2/2026
Show HN: Open-source expense and budget tracker with SQL API for AI agents
Open-source expense tracker with SQL API for AI agents. Five-year transaction dataset, personal finance forecasting features.
3/2/2026
Show HN: A Software-Defined USB PD Charger (ESP32, FPGA, Open Source Hardware)
IFC/BIM building model viewer built entirely in Rust (Leptos, Bevy) compiled to WebAssembly with photometric data visualization.
3/2/2026
OpenPencil – AI-native design editor. Open-source Figma alternative
OpenPencil: open-source, AI-native design editor alternative to Figma. Full read/write MCP support, fully local, active development.
3/2/2026
Show HN: Why LLM Personas Collapse and How Structural Identity Fixes It
Analysis of LLM persona instability in long-running agents. Proposes structural identity layers as architectural solution beyond prompting.
3/2/2026
How to Write a Good Spec for AI Agents
O'Reilly article by Addy Osmani on writing specifications for AI agents. Content appears truncated in provided text.
3/2/2026
Show HN: I built an MCP server to recruit employees for free
MCP server for Himalayas recruiting platform. Enables talent search, evaluation, messaging via AI chat. Job listing and applicant management.
3/2/2026
Show HN: Mini-coder – small, fast CLI coding agent
mini-coder: lightweight terminal-native CLI coding agent built on Bun.js. Fast startup, TypeScript native, minimal UI, mostly self-written.
3/2/2026
X's Rust and Python Architecture Analysed
X's recommendation algorithm rewritten from Scala to Rust+Python microservices architecture for content ranking and serving.
3/2/2026
Show HN: Omni – Open-source workplace search and chat, built on Postgres
Omni: open-source workplace search and chat platform self-hosted on Postgres and pgvector. Connects to Google Drive, Slack, Confluence. Alternative to Glean.
3/2/2026
Let OpenClaw bot to manage your issues and Git repositories
OpenClaw bot framework with Gisia project platform exposing machine-readable skill files enabling AI bots to perform DevOps tasks via REST API.
3/2/2026
Show HN: Deterministic symbolic memory layer for grounding LLMs
Proof-of-concept for deterministic symbolic memory layer accessible via MCP protocol to ground LLMs with explicit knowledge resolution instead of probabilistic recall.
3/2/2026
Show HN: Axiom – structural OCR for handwritten STEM notes
Axiom: structural OCR tool using LLMs to preserve layout and formatting of handwritten STEM notes including equations, derivations, and tables.
3/2/2026
Show HN: OxyJen – Java framework to orchestrate LLMs in a graph-style execution
OxyJen: open-source Java framework for orchestrating multi-step LLM pipelines with graph-style execution, structured type safety, and testability.
3/2/2026
Structured Outputs for LLMs
Implementation of JSON schema structured outputs for LLMs using outlines-core library in custom inference runtime.
3/2/2026
Show HN: Audio-to-Video with LTX-2
LTX-2 open-source diffusion model for audio-to-video generation with prompt enhancement via Gemini for improved synchronization.
3/2/2026
An OpenClaw agent that blogs 24/7 and builds its own host
Long-running autonomous OpenClaw agent managing full blogging pipeline including writing, image generation, git workflows, and deployments.
3/2/2026
Strix – First Impressions
Review of Strix, an agentic AI framework designed for penetration testing and vulnerability discovery.
3/2/2026
Show HN: Steward – an ambient agent that handles low-risk work
Steward ambient AI agent that autonomously handles low-risk work by monitoring GitHub, email, calendar with policy-gated execution.
3/2/2026
Agent with decisions memory at its core
Nous framework for building AI agents using Cognition Engines and Society of Mind principles for decision memory and learning.
3/2/2026
Show HN: Clenv – Manage multiple Claude Code profiles, each Git-versioned
Clenv tool manages multiple Claude Code profiles with git versioning for context isolation across different development roles.
3/2/2026
AI – Factory Model
Analysis of paradigm shift in agentic software engineering where abstraction level changed fundamentally for developers building with AI.
3/2/2026
AI Scientist v3: Scale from 1-hour to 24 hours with Reviewer agent
AI Scientist v3 extends autonomous research from 1-hour to 24-hour experiments using reviewer agent for iterative hypothesis refinement.
3/2/2026
Browser action engine for AI agents. 10× faster, resilient by design
Actionbook provides real-time DOM structure and action manuals to browser automation agents for reliable website interaction without hallucination.
3/2/2026
HumanMCP: A Human-Like Query Dataset for Evaluating MCP Tool Retrieval Performance
HumanMCP dataset providing human-like queries for evaluating tool retrieval performance in Model Context Protocol (MCP) servers.
3/2/2026
An Agentic LLM Framework for Adverse Media Screening in AML Compliance
Agentic LLM system with RAG for automating adverse media screening in AML compliance, reducing false positives from keyword-based approaches.
3/2/2026
MMKG-RDS: Reasoning Data Synthesis via Deep Mining of Multimodal Knowledge Graphs
MMKG-RDS is a framework for synthesizing training data using multimodal knowledge graphs to improve domain model reasoning capabilities.
3/2/2026
PseudoAct: Leveraging Pseudocode Synthesis for Flexible Planning and Action Control in Large Language Model Agents
PseudoAct introduces pseudocode-based planning for LLM agents to reduce token consumption and improve stability on long-horizon multi-tool tasks.
3/2/2026
ODAR: Principled Adaptive Routing for LLM Reasoning via Active Inference
ODAR-Expert presents an adaptive routing framework for optimizing accuracy-efficiency tradeoffs in LLM reasoning via active inference instead of uniform sampling.
3/2/2026
From Flat Logs to Causal Graphs: Hierarchical Failure Attribution for LLM-based Multi-Agent Systems
Method for hierarchical failure attribution in multi-agent LLM systems using causal graphs to improve observability and debugging of complex agent interactions.
3/2/2026
ProductResearch: Training E-Commerce Deep Research Agents via Multi-Agent Synthetic Trajectory Distillation
ProductResearch proposes a multi-agent framework using trajectory distillation to train LLM-based agents for complex e-commerce product research tasks.
3/2/2026
The Auton Agentic AI Framework
Auton framework addresses architectural mismatch between stochastic LLM outputs and deterministic backend systems for agentic AI deployment.
3/2/2026
Unlocking Cognitive Capabilities and Analyzing the Perception-Logic Trade-off
MERaLiON2-Omni: 10B multilingual MLLM for Southeast Asia addressing perception-logic tradeoffs in omni-perception tasks.
3/2/2026
Reasoning-Driven Multimodal LLM for Domain Generalization
Domain generalization method leveraging reasoning chains in MLLMs to improve robustness under domain shift.
3/2/2026
EMO-R3: Reflective Reinforcement Learning for Emotional Reasoning in Multimodal Large Language Models
EMO-R3 applies reflective reinforcement learning to improve emotional reasoning capabilities in multimodal LLMs.
3/2/2026
RUMAD: Reinforcement-Unifying Multi-Agent Debate
RUMAD: reinforcement learning approach to multi-agent debate that adapts topology to task complexity while maintaining debate neutrality.
3/2/2026
RF-Agent: Automated Reward Function Design via Language Agent Tree Search
RF-Agent uses LLM-based tree search to automatically design reward functions for control tasks with improved historical feedback utilization.