3/26/2026
Show HN: Vectimus – Cedar policy enforcement for AI coding agents
Vectimus: Cedar policy enforcement layer for AI coding agents. Blocks dangerous commands and API calls in sub-10ms. Tool for securing agent tool execution.
3/26/2026
Updates to our Privacy Statement and Terms of Service: How we use your data
GitHub updates privacy policy: Copilot Free/Pro users' interaction data used for model training unless opted out. Enterprise unaffected.
3/25/2026
Llama.cpp's Agents.md
llama.cpp project contribution policy restricts AI-generated code, requiring human authorship with AI only for corrections.
3/25/2026
Helix – Self-healing SDK for AI agent payments (open source)
Helix is open-source self-healing SDK enabling AI agents to handle payment transactions with built-in error recovery.
3/25/2026
Attributed Claude activity over the last 90 days on GitHub
Analysis of Claude's GitHub activity over 90 days showing adoption patterns, commits, and code distribution across repositories.
3/25/2026
Tamp – Compression Proxy: 52% Fewer Tokens for Claude Code, Gemini, etc.
Tamp token compression proxy reduces input tokens 52.6% for Claude Code, Aider, Cursor, and compatible coding agents.
3/25/2026
AI-powered pull request evaluation for open source maintainers
AI-powered pull request evaluation tool for open source maintainers.
3/25/2026
The Vibetracer
Vibetracer records real-time AI coding assistant edits with time-scrubbing TUI to inspect diffs and surgically restore files.
3/25/2026
Major conference catches illicit AI use – and rejects papers
ICML 2026 rejected 497 papers for violating AI-use policies in peer review, detected via watermarking system.
3/25/2026
Tamp: Cut LLM context size ~50% without changing your code
Tamp is token compression proxy reducing LLM context size 50% for coding agents without code changes via tool result deduplication.
3/25/2026
Don Cheli – 72 command SDD framework for Claude Code with TDD as iron law
Don Cheli is an open source SDD framework for Claude Code with TDD, automatic complexity detection, Gherkin specs, and Docker execution.
3/25/2026
Show HN: Agent Kernel – Three Markdown files that make any AI agent stateful
Agent Kernel: stateful AI agents using three Markdown files and git for memory across sessions.
3/25/2026
Grove: Distributed ML Training over AirDrop
Grove enables distributed ML training across MacBooks via AirDrop without network configuration.
3/25/2026
Poking Around ChatGPT's Sandbox
Technical exploration of ChatGPT's code execution sandbox, documenting filesystem access, networking, and process limitations.
3/25/2026
Show HN: A Technique for Self-Improving Agents
Orchestrator agent pattern that creates isolated sub-agent clones in git worktrees to convert websites to JSON APIs with self-improvement analysis.
3/25/2026
Your AI Agent Has a Memory Problem
Title only, no content provided. Discusses memory issues in AI agents.
3/25/2026
Show HN: MonkePay – Charge AI Agents per API Request in USDC
MonkePay: middleware wrapper for x402 protocol enabling per-request USDC payments for AI agent API access.
3/25/2026
Show HN: First autonomous AI agent purchase via open protocol (UCP)
First autonomous AI agent purchase via Universal Commerce Protocol completed on WooCommerce without human intervention.
3/25/2026
Kern – AI agents built for coworking
Kern: multi-channel AI agents for team collaboration across Slack, Telegram, and terminal with unified memory.
3/25/2026
Tmux Replacement for Agentic Workflows
Terminal multiplexer and orchestrator for multi-agent Claude Code sessions with PTY-backed panes and IPC messaging.
3/25/2026
Uploading to Wikimedia Commons with AI
AI tool automates categorization when uploading photos to Wikimedia Commons.
3/25/2026
AutoGaze removes redundant video patches before passing into a ViT or MLLM
AutoGaze reduces video patch redundancy for vision transformers via reinforcement learning, achieving 4-100× token reduction.
3/25/2026
Bruno: The Git-native API client
Open source Git-native API client with collections stored as code, designed for developers and agents.
3/25/2026
AI and Git Worktree: a multi-agent setup that works
Using git worktree to run multiple AI coding agents in parallel on complex codebases.
3/25/2026
Show HN: Druids – coordinate and deploy coding agents across machines
Druids library coordinates and deploys coding agents across machines using Python functions and event-driven architecture.
3/25/2026
My story on how I created my own open source project
Open source tool called 'why' that explains C/C++ compiler errors to beginners with suggested fixes.
3/25/2026
Constructing an LLM-Computer
Incomplete article about constructing LLM systems using transformer weights.
3/25/2026
5 AIs. One Answer on iOS
iOS app querying five AI models simultaneously to compare answers and reduce single-source bias.
3/25/2026
Show HN: Rick – Open-source AI CEO that autonomously runs your startup
Rick: open-source autonomous AI agent that manages startup operations independently.
3/25/2026
Specula: A framework for finding deep bugs in system code using TLA+
AI-powered framework using LLMs and TLA+ formal specification to find bugs in distributed system code through agentic workflows for code analysis and trace validation.
3/25/2026
Silence is the KPI: agents that don't talk
Framework for building autonomous AI agents that prioritize shipping outcomes over conversation, with GitHub examples and operating system design principles.
3/25/2026
RedSwarm Adversarial AI security scanner, one file, zero deps
RedSwarm spawns adversarial AI agents to simulate attacks on codebases, detecting supply chain attacks, prompt injection, and credential theft. Single Python file, zero dependencies.
3/25/2026
Show HN: Whorl – Fingerprinting LLMs as horrible password generators
Research on LLM fingerprinting via password generation patterns. Tool achieves 92% model identification accuracy from 5 samples. Demonstrates model-specific text patterns.
3/25/2026
Joycraft: Upgrade your Claude Code harness
CLI tool and Claude Code plugin that structures AI-assisted development workflows with skills, templates and spec-driven autonomous development patterns.
3/25/2026
How to make good lecture slides with AI assistance
Blog post about using Claude to assist with creating university lecture slides and course materials.
3/25/2026
Show HN: GhostDesk – MCP server giving AI agents a full virtual Linux desktop
MCP server providing LLM agents with Linux desktop UI control via mouse/keyboard/screenshots. Enables agents to operate software and websites like humans.
3/25/2026
Show HN: LiveDemo – open-source tool for creating interactive product demos
Open-source tool for auto-generating interactive product demos without video recording. Self-hostable with automatic demo flow generation.
3/25/2026
Launch: A terminal process manager to run your multiple-service project
Terminal process manager for running multiple services locally with unified logging and control. Open source developer tool with YAML configuration.
3/25/2026
Agents can write dlt pipelines. Now they can run and deploy them
dltHub AI Workbench infrastructure enabling AI agents to autonomously write, run and deploy production data pipelines with 81k+ custom sources created monthly.
3/25/2026
AI Efficiency Courses
Comprehensive online course covering building, compressing, evaluating, and deploying efficient AI models.
3/25/2026
Show HN: An offline first focused agentic CLI application powered by Ollama
Offline-first agentic CLI application using Ollama local models with advanced memory management; MIT licensed.
3/25/2026
Blatant-Why: Stop paying for AI agent wrappers, start asking what's underneath
Open-source AI agent framework with 11 MCP servers, 16 agents, and Claude integration; eliminates platform fees.
3/25/2026
Updates to GitHub Copilot interaction data usage policy
GitHub updates Copilot data usage policy to train AI models on user interactions (code, inputs, outputs) starting April 24, with opt-out option.
3/25/2026
Construction project manager takes on software development
Developer transitioning from carpentry to building AI agents, launching first product.
3/25/2026
ImpossibleBench: Measuring LLMs' Propensity of Exploiting Test Cases
Research on measuring whether LLMs exploit patterns in test cases rather than genuine understanding.
3/25/2026
Show HN: Got a tough conversation coming up? This AI app will help you prepare
PostHog shares lessons learned from building AI agents in production environments.
3/25/2026
Claude Code gets 'safer' auto mode
Update to Claude Code with safer auto-execution mode for code generation.
3/25/2026
Updates to GitHub Copilot interaction data usage policy
GitHub announces Copilot interaction data will be used for model training unless users opt out.
3/25/2026
Google's TurboQuant offers LLMs up to 6x compression
Google's TurboQuant compression algorithm reduces LLM memory footprint up to 6x while maintaining output quality.
3/25/2026
Show HN: Pgsemantic – Point at your Postgres DB, get vector search instantly
Tool adding vector semantic search to PostgreSQL databases with MCP integration for AI agents.