2/16/2026
CodeForge – 100 AI agents review your code like hostile attackers
AI agent system using 100 LLM instances to review code from adversarial perspective for security testing.
2/16/2026
CURL's Daniel Stenberg: AI slop is DDoSing open source
CURL maintainer discusses AI-generated content causing performance issues for open source projects.
2/16/2026
Interface-Off: Which LLM designs the best marketing site?
NadirClaw: open-source LLM router with 10ms prompt classification. Routes to cheap/premium models automatically via OpenAI-compatible proxy.
2/16/2026
AgentDocks – open-source GUI for AI agents that work on your real codebase
Open-source GUI enabling AI agents to autonomously work on and modify real codebases with visual interface.
2/16/2026
Security audit for LLM skill files: skillaudit.sh
Tool for security auditing LLM skill files and agent configuration files. Automated vulnerability detection for AI agent setups.
2/16/2026
The Long Tail of LLM-Assisted Decompilation
LLM-assisted decompilation of Nintendo 64 games using coding agents, achieving 75% code match for Snowboard Kids 2 with iterative workflow refinement.
2/16/2026
BeadHub: Allow coding agents to claim work, chat, and coordinate across machines
Coordination framework enabling multiple AI agents to claim tasks, communicate, and collaborate across distributed machines.
2/16/2026
Don't Trust the Salt: AI Summarization, Multilingual Safety, and LLM Guardrails
Technical work on LLM safety via multilingual guardrails and summarization. Research on AI safety mechanisms.
2/16/2026
Show HN: SecureClaw – Open-Source Security Layer for OpenClaw Agents
Open-source security layer for OpenClaw agents. Developer tool for securing AI agent operations.
2/16/2026
Guardian: Role-Gated MPC Wallets for AI Agents
Guardian system providing MPC wallet infrastructure for AI agents. Security and blockchain integration for autonomous agents.
2/16/2026
[Android]Nabu 0.5.4 – supporting Soprano TTS and local LLM HTTP server
Nabu 0.5.4: Android app with local LLM HTTP server support and Soprano TTS integration.
2/16/2026
Gobii vs. OpenClaw: Timeline, Architecture, and Always-On Agents
Comparison of Gobii vs. OpenClaw agent architectures and always-on capabilities. Technical overview of agent systems.
2/16/2026
Turning Production Logs into Evaluation Datasets: A Data-Driven Approach
Technical approach for generating evaluation datasets from production logs for LLM/ML model evaluation.
2/16/2026
Show HN: Math-spec-driven LLM skill for complex system specs
LLM agent skill using math-style specs (sets, relations, invariants) to generate architectures, APIs, and tests for complex system specs.
2/16/2026
Show HN: AsdPrompt – Vimium-style keyboard navigation for AI chat responses
Chrome extension adding Vimium-style keyboard navigation to AI chat interfaces for faster text selection and navigation.
2/16/2026
Show HN: Synrix local-first memory engine (O(k) retrieval, no vectors, no cloud)
Local-first memory engine for AI systems with O(k) retrieval, no vectors, no cloud dependency. Open source developer tool.
2/16/2026
Show HN: Claude-engram – Brain-inspired persistent memory, runs inside Claude.ai
Memory system for Claude.ai inspired by human cognition including salience scoring and forgetting curves, runs as React artifact.
2/16/2026
Show HN: Add functional custom buttons/panels to any site via prompts
Chrome extension using LLMs to generate custom UI components and integrations based on natural language prompts.
2/16/2026
Show HN: DroidClaw – Turn old Android phones into AI agents
Open-source tool converting plain English instructions into automated Android phone tasks via LLM and accessibility tree parsing.
2/16/2026
GenAI does not just hallucinate at us, it can hallucinate with us, study warns
Research study on how human-AI interactions can lead to shared hallucinations and inaccurate beliefs.
2/16/2026
Show HN: CabbageSEO: Check if AI mentions your business, then fix it if not
CabbageSEO: Tool scanning AI platforms for business mentions, scoring visibility, and suggesting SEO fixes.
2/16/2026
Write your own LOTR story with LORA BLOOM-3B (2023)
LORA fine-tuned BLOOM-3B model generating Lord of the Rings-style stories; deployment unavailable for testing.
2/16/2026
Semantic Firewall v3: A Practical Audit Layer for AI
Semantic Firewall v3: Audit layer for AI systems that pre-classifies inputs to reduce hallucination and compute waste before reaching large models.
2/16/2026
Show HN: Multi-provider iOS usage alerts for AI subscription usage caps
AI Usage Tracker: iOS app monitoring multiple LLM subscription usage (OpenAI, Anthropic, etc.) with alerts and widgets for usage caps.
2/16/2026
Maestro – Autonomous dev platform with multi-reviewer AI safety pipeline
Maestro: Autonomous development platform with multi-reviewer AI safety mechanisms for code generation.
2/16/2026
Which AI deep research agent is the current best?
Discussion/question comparing current AI deep research agent capabilities without detailed analysis.
2/16/2026
Show HN: MCP server that gives AI agents real interactive terminal sessions
MCP server enabling AI agents to execute interactive terminal sessions with persistent state and real-time feedback.
2/16/2026
GGML GGUF File Format Vulnerabilities
Security vulnerabilities in GGUF file format used by GGML library for model weights. Technical security research.
2/16/2026
The OpenClaw Security Guide No One Wants to Write (But Everyone Needs)
Security analysis of deploying OpenClaw agents for infrastructure management. DevOps perspective on AI agent risks.
2/16/2026
My AI now has a license to drive
Autonomo: Tool enabling AI assistants to interact with UIs, see application state, and verify changes on real devices instead of blind code suggestions.
2/16/2026
Show HN: A Claude meta-skill that improves all your skills, including itself
Open-source Claude skill that observes work sessions, captures corrections, and automatically improves skills through feedback. Domain-agnostic meta-learning system.
2/16/2026
Project Vesta habitat where AI agents breed, evolve, and run experiments
Project Vesta: Environment for AI agents to breed, evolve, and conduct experiments autonomously.
2/16/2026
Show HN: API authentication assumes the caller decides – agents break that
Analysis of authentication gap in LLM agents: APIs authenticate the process, not the decision-maker (the model). Explores security implications of prompt injection and misalignment.
2/16/2026
Clean: High Performance Prompt Injection Detection and Mitigation
Fast, span-level prompt injection detection and mitigation for AI agents. No GPU or API calls required. Detects attacks in emails, CSVs, webpages, and other external content sources.
2/16/2026
Show HN: A joystick for your LLM/Agent
Interactive tool for mid-generation prompt steering via joystick interface. Drag-and-drop prompt component mixing with pause/resume generation control.
2/16/2026
Show HN: API router that picks the cheapest model that fits each query
Komilion: API router that classifies queries and routes to cheapest suitable model from ~390 options. Drop-in OpenAI SDK replacement with three quality tiers.
2/16/2026
The Benchmark Gap: What It Takes to Ship Kimi K2.5
Article about Kimi K2.5 LLM shipping decisions and benchmark gaps. Title only, minimal content.
2/16/2026
Ask HN: In a blind coding test, could you identify an LLM strictly off vibes?
Discussion question: can developers identify LLM models by coding output characteristics without knowing which model was used?
2/16/2026
Show HN: Regression tests for detecting cross-domain hallucinations in LLMs
Regression test dataset (40 edge cases) for detecting cross-domain hallucinations in LLMs, covering CVE-regulation bindings and tech-legal causality failures.
2/16/2026
Show HN: Agent-history project-wide full-text search for Codex/Claude logs
Full-text search TUI for Claude and Codex conversation logs with real-time indexing and no external dependencies.
2/16/2026
Show HN: A "content compiler" that turns LLM output into validated artifacts
Tool that compiles LLM output into validated, shippable artifacts. Enforces structural invariants to ensure generated content is complete and usable.
2/16/2026
Show HN: Axel – native macOS app with a Rust CLI to orchestrate agents
Axel is a native macOS app with Rust CLI for orchestrating multiple AI agents across terminals and git worktrees, featuring a Things-inspired UI. Currently in beta.
2/16/2026
Why LLMs keep editing the wrong file
Analysis of why LLMs mislocalize and edit wrong files with multiple sources of truth. Explores LLM Experience (LX) architecture optimizing for machine collaboration.
2/16/2026
Show HN: Claude Battery – usage at a glance. A minimalist macOS menu bar widget
Minimalist macOS menu bar widget to track Claude API usage at a glance. Helps non-technical users monitor LLM costs and session limits.
2/16/2026
Show HN: J-RAY – A privacy-first JSON visualizer with two-way binding
Open-source, client-side JSON visualizer with two-way binding and interactive nested data exploration. Privacy-first alternative to online formatters.
2/16/2026
Show HN: Codex HUD – Claude-HUD Style Status Line for Codex CLI
Open-source terminal UI for monitoring Codex CLI sessions with real-time model, project, and usage tracking.
2/16/2026
Show HN: Kai – A Telegram bot that turns Claude Code into a personal dev asst
Kai: Telegram bot wrapper for Claude Code providing shell access, file editing, git, and web search from mobile.
2/16/2026
Sourdine – open-source macOS app for meeting transcription with 100% local AI
Sourdine: Open-source macOS app for local AI-powered meeting transcription with 100% on-device processing.
2/16/2026
Show HN: Klovr – Convert any webpage to Markdown (Cloudflare covers only 5%)
Klovr: Tool converting webpages to Markdown format. Utility tool with limited AI relevance.
2/16/2026
Show HN: InitRunner – YAML to AI Agent with RAG, Memory, and an API
InitRunner: YAML-based platform for rapid AI agent prototyping with RAG, memory, and API endpoints.