6/11/2026
Hacking Salesforce Sites with an LLM Agent
Demonstrates using LLM agents to automate Salesforce Site interactions, showing practical LLM application for enterprise system automation.
6/11/2026
Show HN: Flightdeck – self-hosted observability and control for AI agents
Self-hosted observability and control platform for AI agents. Provides monitoring, debugging, and management capabilities for agent systems.
6/11/2026
Show HN: Fabrika, from babysitting agents to software factories
Fabrika orchestrates multiple CLI coding agents in parallel across isolated git worktrees with test validation before merging, converting natural language tasks to production code.
6/11/2026
Show HN: Lumen–free Real-time LLM token and cost monitor
Free LLM token and cost monitoring tool for real-time tracking of LLM usage.
6/11/2026
ZenNotes – Open-source local Markdown app with system-wide Vim motions
Keyboard-first Markdown notes app with Vim motions, local storage, MCP integration, multi-platform support.
6/11/2026
Show HN: 5dive – Run a Team of Claude Code Agents from Telegram
Run team of Claude Code agents from Telegram, multi-agent coordination with MIT license, self-hostable.
6/11/2026
Show HN: Editxr – open-source WYSIWYG Markdown editor for the terminal
Terminal-based WYSIWYG Markdown editor in Swift with AI-assisted editing and inline diffs, zero dependencies.
6/11/2026
Kasspian – An AI that tells you your business idea isn't good
AI application that provides critical feedback on business ideas.
6/11/2026
Open Reproduction of DeepSeek-R1
Open-source reproduction of DeepSeek-R1 reasoning model pipeline, community-driven implementation effort.
6/11/2026
Flatiron is a fast columnar analytics library for Clojure
Fast columnar analytics library for Clojure with SQL-like DSL and graph algorithms, pure Clojure implementation.
6/11/2026
Show HN: Run production AI in your cloud in 5 mins
Dagploy open-source tool for self-hosted AI model deployment and infrastructure management in 5 minutes.
6/11/2026
You're Doing AI Agents Wrong – Engineer the Harness
Article on engineering best practices for building AI agents effectively.
6/11/2026
TCS ties up with Anthropic to roll out Claude access to 50k employees
TCS partnership with Anthropic to provide Claude LLM access to 50,000 employees for enterprise adoption.
6/11/2026
Show HN: A better LLM-wiki with multi-path research [550 stars]
LLM-powered wiki with multi-path research. Limited details in stub.
6/11/2026
Are we defaulting to VM-level sandboxing before understanding the threat model?
Examines sandboxing approaches for running model-generated code, questioning whether VM-level isolation is always necessary vs threat model requirements.
6/11/2026
Anthropic's AI Jobs Paper
Anthropic research paper on AI's economic impact on employment and job markets.
6/11/2026
Outpost – Capability-based API access for AI agents
Outpost provides capability-based API access control for AI agents, enabling secure tool use.
6/11/2026
Malware devs added text to trigger LLM safety refusal, to avoid detection
Security research: malware developers exploit LLM safety refusal mechanisms by injecting trigger text to evade detection.
6/11/2026
Anthropic Walks Back Policy That Could Sabotage AI Researchers Using Claude
Anthropic reverses Claude safeguards policy that restricted AI researchers' capability testing.
6/11/2026
Show HN: SpadeBox – Sandboxed tools and JavaScript runtime for AI agents
SpadeBox provides sandboxed JavaScript runtime and tool environment for autonomous AI agents.
6/11/2026
Markov Chains: The Grandparents of LLMs
Conceptual article connecting Markov chains as foundational mathematical concepts underlying LLMs.
6/11/2026
Making a Vintage LLM from Scratch
Article on building an LLM from scratch, likely educational overview of transformer architecture fundamentals.
6/11/2026
Evaluating LLM-generated code for domain-specific languages
Research evaluating LLM-generated code quality specifically for domain-specific languages.
6/11/2026
Ask HN: Temporal Awareness in LLM?
Discussion thread requesting temporal awareness feature for LLMs to track elapsed time in conversations.
6/11/2026
Run OSINT investigations from chat via MCP/OpenClaw
Self-hosted OSINT toolkit accessible via MCP server, CLI, web app, and chat apps (WhatsApp/Discord/Telegram).
6/11/2026
Discussion: Fable 5 is weak at flagging prompts correctly
Discussion of Fable 5 language model's weak prompt flagging behavior and limitations.
6/11/2026
Open-source Linux tray app for tracking Claude Code and Codex usage
OpenUsage Community: cross-platform Linux tray app for tracking Claude and Codex AI coding subscriptions.
6/11/2026
Hate using Granola? Here is a free local version
Hobnob: local meeting transcription and summarization tool using Whisper and bundled LLM, no cloud.
6/11/2026
CommBench: Can LLMs Write Correct and Efficient GPU Communication Code?
Benchmark evaluating LLM capability to write correct and efficient GPU communication code.
6/11/2026
Frontier: A Discrete-Event Simulator for Modern LLM Serving
Discrete-event simulator for evaluating LLM serving systems and inference optimization.
6/11/2026
AI researcher claims he's bypassed Anthropic's Fable 5 guardrails
Researcher claims to have jailbroken Claude Fable 5 within 48 hours of launch. Security/safety analysis without technical details.
6/11/2026
Is It AI? How to Tell Using Metadata
Article on detecting AI-generated images using metadata and forensic techniques. Addresses misinformation risk from generative content.
6/11/2026
Show HN: A command-line story machine inspired by Roald Dahl's 1953 short story
Command-line story generation tool using OpenAI-compatible APIs and local models. Based on Roald Dahl concept, works with any LLM endpoint.
6/11/2026
Validation, Docs, tests, and database schemas from one source of truth
TypeScript framework generating validation, OpenAPI, tests, and database schemas from single source definition. Developer tooling for API specification and implementation.
6/11/2026
Show HN: Vaportrail – flight recorder for Claude Code, Codex, and OpenCode
Flight recorder tool that captures and analyzes transcripts from Claude Code, Codex, OpenCode, and other AI coding agent sessions. Searchable local-first history.
6/11/2026
Grit by GitButler: A from-scratch reimplementation of Git in idiomatic Rust
Git reimplementation in Rust built entirely by AI coding agents, passing C Git test suite. Demonstrates agent capabilities in substantial software engineering.
6/11/2026
Inverse Rubric Optimization: A testbed for agent science
Research on inverse rubric optimization as testbed for studying long-horizon agent performance and general principles of agent science.
6/11/2026
Ask HN: Releasing code under AGPLv3, but want to block LLM reconstruction?
Discussion on using AGPLv3 licensing to prevent LLM training ingestion and closed-source reconstruction of open-source code.
6/11/2026
Show HN: Headroom – Claude Code usage limits, live in the macOS menu bar
macOS menu bar app monitoring Claude Code API usage limits in real-time with session and weekly quota tracking.
6/11/2026
Auto-geo – open-source CLI for GEO that helps get your brand mentioned by LLMs
Open-source CLI tool to audit and track which domains LLMs cite in responses, with file-based tracking and support for ChatGPT, Claude, Gemini.
6/11/2026
OpenAI to acquire Ona
OpenAI acquires Ona to integrate secure cloud execution and orchestration into Codex ecosystem for long-running AI agents. Codex usage grew 400% to 5M weekly users.
6/10/2026
GGUF vs. GPTQ vs. AWQ: The Plain-English Guide to LLM Quantization
Guide to LLM quantization formats (GGUF, GPTQ, AWQ) with VRAM calculations and decision framework for running large models locally.
6/10/2026
Agent-gate – fail-closed agent gate and tamper-evident receipts as an MCP server
MCP server implementing fail-closed gating and tamper-evident receipts for AI agents with deterministic checks and independent review.
6/10/2026
Loop-Harness
Autonomous loop harness for coding agents using Claude with task-specific skills, isolated git worktrees, and verification before shipping output.
6/10/2026
Show HN: I let an AI C-suite run my company – starter kit from the inside
Startup kit documentation for running autonomous AI agents in C-suite roles with system prompts and agent configurations.
6/10/2026
Show HN: Llmbuffer – Python library for cache-optimized LLM conversation history
Python library optimizing LLM conversation history caching for agent threads with dynamic context and flexible hooks.
6/10/2026
/dmg – a Claude Code skill for persistent memory and session sync
Claude Code slash command maintaining persistent memory across sessions via documentation, memory system, and git commits.
6/10/2026
Nuts – pip/NPM for Java with first-class workspaces and JDK provisioning (9y+)
Nuts package manager for Java providing workspace support and JDK provisioning, addressing distribution challenges.
6/10/2026
Show HN: Magenta Real-Time Music Generation on iPhone, Without the GPU
Developer optimized DeepMind's Magenta music generation model to run on iPhone 12 Pro without GPU, demonstrating on-device ML inference techniques.
6/10/2026
Propel: Breaking the Solver Bottleneck in Task-Generator RL
Propel addresses task-generator bottleneck in RL by training systems to generate harder, discriminative tasks as agent capabilities improve, moving beyond fixed benchmarks.