3/13/2026
Show HN: Tarvos – Coding agents that work infinitely
Tarvos: Coding agent using Relay architecture to cascade new agents before context degradation, enabling indefinite coding tasks.
3/13/2026
Unified Sandbox SDK for AI Agents
SDK for building and testing AI agents in sandboxed environments. Limited details on implementation.
3/13/2026
My AI Agent Knows Who I Am. Not Just What I Want. Who I Am
Trading system using Claude with persistent memory and tool-use architecture. Addresses LLM limitations in agent design.
3/13/2026
The AI productivity paradox: More work, not less
AI productivity gains redirect to increased workload rather than time savings. Analysis of corporate AI adoption creating more work for employees.
3/13/2026
Building a strong data infrastructure for AI agent success
Enterprise data infrastructure for AI agent deployment. Sponsored content on building systems for agentic AI in enterprises.
3/13/2026
MaximusLLM: High-Speed Architecture via Ghost Logits and Random Latent Attention
Novel LLM architecture decoupling computational cost from sequence length using RandNLA Attention, MAXIS Loss, and Fisher-SVD for long-context modeling.
3/13/2026
Prompt-caching – auto-injects Anthropic cache breakpoints (90% token savings)
Prompt caching plugin for Claude auto-detects stable content and applies Anthropic cache breakpoints, reducing token costs by 90%.
3/13/2026
Western AI models "fail spectacularly" in farms abroad
AI models fail on local crops in Kenya without local data adaptation. Study shows Western AI lacks generalization to non-Western agricultural conditions.
3/13/2026
Show HN: RAG Doctor – CLI tool to diagnose broken RAG pipelines
CLI tool for diagnosing RAG pipeline failures by identifying chunking, embedding, retrieval, and injection vulnerabilities.
3/13/2026
Want to Win a New CanaKit Raspberry Pi 5 Starter Kit Pro?
Open-source PostgreSQL MCP server with prompt caching, token efficiency, and AI-powered chat for database interactions.
3/13/2026
Prompt engineering vs. context engineering: a practical guide for AI builders
Guide contrasting prompt engineering vs context engineering for LLM applications, addressing real-world pilot failures.
3/13/2026
Analysis → Implementation → Reflection – a practical technique for agentic AI
Describes Analysis/Implementation/Reflection pattern for agentic AI with exploration harness and qualitative assessment phases.
3/13/2026
Show HN: Wardstone – Prompt injection and jailbreak detection API
Wardstone API for detecting prompt injection attacks, jailbreaks, and harmful content in LLM inputs/outputs with sub-30ms latency.
3/13/2026
Show HN: Who watches the watchmen? A public decision track record for AI agents
Agent Smith tool for tracking AI agent decision-making records with confidence scores and outcome auditing.
3/13/2026
The Controllability Trap: A Governance Framework for Military AI Agents
Research framework examining governance and controllability challenges in military AI agents.
3/13/2026
I mass-replaced FFmpeg's MJPEG decoder with Claude Code – 4K LOC, 8% the speed
AI-generated MJPEG decoder written by Claude Code in pure C99. Performance comparison shows FFmpeg 12x faster due to hand-tuned SIMD assembly.
3/13/2026
Need feedback to build a product for founders to track decision-making
Proposal for AI-powered decision-tracking chatbot for founders to record reasoning and improve decision-making.
3/13/2026
I got tired of AI chatbots so we turned the OS into an AI agent
OS-integrated AI agent replacing traditional chatbots for system interaction. Project overview with implementation approach.
3/13/2026
Show HN: Privacy Mask – prevent secrets leaking to AI agents
Open-source tool intercepting and redacting screenshots locally to prevent accidental secret leaks to AI agents.
3/13/2026
Beyond Agents.md: Harness Eng, Loop-Based Delivery, and Context-Aware Prompting
Technical guide on loop-based agentic workflows, harness engineering, and context-aware prompting for production codebases.
3/13/2026
Show HN: Agile V Skills – Open skills for verifiable, traceable AI engineering
Agile V Skills framework for verifiable, traceable AI agent software engineering with independent testing and requirement traceability.
3/13/2026
Rust and LLMs: The Compiler Does What Code Review Shouldn't Have To
Using Rust with AI coding tools; discusses compiler benefits and learning curve reduction for developers.
3/13/2026
Blader/humanizer: Claude Code skill to remove AI-generated tells from writing
Claude Code skill that removes AI-generated writing patterns from text to produce more natural-sounding output.
3/13/2026
Ask Maps and Immersive Navigation: New AI Features in Google Maps
Google Maps integrates Gemini models for conversational location queries and immersive navigation features.
3/13/2026
Show HN: CacheLens – Local-first cost tracking proxy for LLM APIs
CacheLens local proxy tracks LLM API costs, token usage, cache hit rates, and latency across providers with real-time dashboard.
3/13/2026
AutoExp: One-liner turn any training code to autoresearch
AutoExp: One-line setup turning ML training code into automated research workflows using AI coding agents to optimize experiments iteratively.
3/13/2026
What if compiler errors were an API? (AI-native language demo)
Valea: minimal systems language with JSON-based compiler API designed for AI agent code generation without error scraping.
3/13/2026
Ask HN: What benchmarks do you trust most when comparing large LLMs?
Discussion thread asking community which LLM benchmarks (IFBench, SWE-Bench, Tau Bench, RULER) are most trusted for real-world model evaluation.
3/13/2026
LLMs: Using a single Unix-style tool instead of multiple tools/function calling
Discussion comparing Unix-style tool approach versus function calling for LLM agent implementations based on 2 years production experience.
3/13/2026
Ask HN: Resources for a conceptual model of LLMs as applicable to coding?
Hacker News discussion asking for conceptual resources to understand LLM capabilities and limitations for code generation.
3/13/2026
Same Chat App, 4 Frameworks: Pydantic AI vs. LangChain vs. LangGraph vs. CrewAI
Full-stack template comparing 4 AI frameworks (Pydantic AI, LangChain, LangGraph, CrewAI) with identical chat application implementation across all.
3/13/2026
Agentic, fully-automated reverse engineering
Kong: LLM-orchestrated agent for automated binary reverse engineering using NSA-grade frameworks to analyze obfuscated binaries.
3/13/2026
Show HN: From Claude Code to OpenCode – My Evolution in Vibe AI Engineering
Personal documentation of transitioning from Claude Code to open-source OpenCode stack for AI-assisted development workflows.
3/13/2026
ShowHN: Turn PDFs, notes and spreadsheets into business briefs
Gixo AI tool for converting PDFs, notes and spreadsheets into business briefs with structured templates and collaborative editing.
3/13/2026
Show HN: SiMM – Distributed KV Cache for the Long-Context and Agent Era
SiMM distributed KV cache system addressing long-context LLM inference bottlenecks with reduced GPU memory requirements and faster time-to-first-token.
3/13/2026
Crazy Rogue AI
Provisional patent application for cryptographically accountable multi-agent AI pipeline architecture with structural safety enforcement and bi-directional logging.
3/13/2026
MCP Doesn't "Suck"
Defense of Model Context Protocol (MCP) against criticism about context-window bloat and authentication. Argues MCP is flexible protocol, not implementation problem.
3/13/2026
Show HN: Adversarial Code Review paired agents, zero noise,validated findings
Code review system using two opposed AI agents: reviewer finds problems, dev agent disproves findings. Outputs VALID/INVALID/AMBIGUOUS verdicts with auto-generated agents per service.
3/13/2026
How we hire AI-native engineers now: our criteria
Analysis of how companies are adapting hiring practices as AI agents write most code. Focus shifts from implementation ability to product taste and architectural judgment.
3/13/2026
Agent Engine Optimization (AEO): Selling to AI Agents
Claude Code skill using Subconscious Systems agent for agent engine optimization (AEO). Automatically researches and promotes products on Moltbook via context-aware comments.
3/13/2026
AI-generated passwords aren't random, it just looks that way
Research shows Claude, ChatGPT, and Gemini generate weak passwords despite appearing complex, failing random-ness requirements.
3/13/2026
Show HN: ClawRemove – Inspect and clean AI agent environments
ClawRemove is an agent environment inspector tool for auditing and cleaning runtime environments where AI agents execute.
3/13/2026
Show HN: Droeftoeter, a Terminal Coding Toy
Droeftoeter terminal toy lets users prompt an LLM to generate and extend ASCII art animations on a 64x32 grid.
3/13/2026
CLI-Anything
CLI-Anything framework converts software into agent-ready interfaces via structured command-line protocols, enabling AI agents to interact with legacy systems.
3/13/2026
Hybrid Human-Agent Social Dilemmas in Energy Markets
Studies emergence of cooperative behavior in hybrid human-agent populations through energy load management game-theoretic scenario.
3/13/2026
The Landscape of Generative AI in Information Systems: A Synthesis of Secondary Reviews and Research Agendas
Systematic review synthesizing 28 secondary studies on generative AI adoption in organizations, identifying technical and organizational challenges.
3/13/2026
You Told Me to Do It: Measuring Instructional Text-induced Private Data Leakage in LLM Agents
Identifies Trusted Executor Dilemma: high-privilege LLM agents executing external instructions leak private data through instruction-following vulnerability.
3/13/2026
ELISA: An Interpretable Hybrid Generative AI Agent for Expression-Grounded Discovery in Single-Cell Genomics
ELISA integrates scGPT expression embeddings with BioBERT for interpretable AI agent discovery in single-cell genomics data.
3/13/2026
The Mirror Design Pattern: Strict Data Geometry over Model Scale for Prompt Injection Detection
Mirror design pattern for prompt injection detection using data-curation geometry for fast, deterministic, non-promptable screening.
3/13/2026
Bielik-Minitron-7B: Compressing Large Language Models via Structured Pruning and Knowledge Distillation for the Polish Language
Bielik-Minitron-7B compresses Bielik-11B model 33.4% via structured pruning and knowledge distillation for European languages.