AI Watchtower Briefing — 2026-07-01

🔴 High Significance

Model Releases

🔴 💬 Non Us Ally should be afraid. — score 82 Sources: reddit/r/LocalLLaMA

Spyware-like code in Claude Code that covertly targets Chinese users.

Developer Tools

🔴 💬 Looking for feedback from people building AI agents that use browsers — score 94 Sources: reddit/r/AIAgents

I'm interested in getting technical feedback from people building AI agents that interact with websites. One issue I kept running into was Playwright sessions getting identified by anti-bot systems, so I started experimenting with a Firefox fork that modifies fingerprinting at the browser level inst

🔴 💬 How do you use AI agent working for social media data analysis? — score 83 Sources: reddit/r/AIAgents

Just want to know which ai agent can do or how to use these diverse agents to do analysis on social media platforms and give insigts on content creation?

🔴 🐙 virgiliojr94/book-to-skill — Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work. — score 76 Sources: github_trending

Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work.

🔴 🐙 open-webui/open-webui — User-friendly AI Interface (Supports Ollama, OpenAI API, ...) — score 70 Sources: github_trending

User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

Infrastructure & Compute

🔴 🐙 allenai/olmocr — Toolkit for linearizing PDFs for LLM datasets/training — score 80 Sources: github_trending

Toolkit for linearizing PDFs for LLM datasets/training

Business & Funding

🔴 💬 On July 1, 2026, arXiv will spin out from Cornell University, its home for the past 25 years, to become an independent nonprofit organization. Major funding support from Simons Foundation and Schmidt Sciences. Ditching the red for their website. [N] — score 94 Sources: reddit/r/MachineLearning

arXiv’s next chapter: Updates on our spin out from Cornell University: https://blog.arxiv.org/2026/06/30/arxivs-next-chapter/

Other Signals

🔴 💬 The gap between closed and open models might be much smaller than commonly assumed, because we don’t know what closed model providers do in addition to model inference — score 96 Sources: reddit/r/LocalLLaMA

When Claude dominates GLM-5.2 in benchmarks, it’s usually assumed that Anthropic has superior model architectures, superior training pipelines, and other advanced machine learning techniques that make their models better than the competition. But actually, this doesn’t follow. Because the benchmarks

🔴 💬 Couldn't hold back — score 89 Sources: reddit/r/LocalLLaMA

Had been waiting for months and the cards finally got delivered today. No one at my workplace was excited, maybe because no one cares for AI stuff that i work on. But I just wanted to share it with you guys. Can't wait to build the server and start working on them.

🔴 💬 SWE-rebench leaderboard update: GLM-5.2, Qwen3.6-27B, Qwen3.6-35B-A3B, Gemma 4 31B and more + improved UI — score 75 Sources: reddit/r/LocalLLaMA

Hi all, We made several updates to the SWE-rebench leaderboard: added new models, refreshed recent results, and reworked the leaderboard UI to make results easier to read, compare, and understand. New Models: * Claude Opus 4.8 xhigh: 56.5% — 2.48M tokens * GLM-5.2: 51.1% — 2.62M tokens * Gemini 3.5

🟡 Notable

Model Releases

🟡 💬 Open Models - June 2026 — score 61 Sources: reddit/r/LocalLLaMA

After overwhelming April, OK May, here's June. Yeah, Graph has only less items. Because we got other items here last month. Finetunes: * Nex-N2 * Ornith-1.0 * Agents-A1 * Holo3.1 * Tmax-27b *

🟡 𝕏 @GoogleDeepMind: We’re shipping 2 major releases: 🔘 Nano Banana 2 Lite: our fastest and cheapest Gemini Image model 🔘 Gemini Omni Flash: now available via the Gemini API and in @GoogleAIStudio to help developers gene — score 60 Sources: twitter_rss

We’re shipping 2 major releases: 🔘 Nano Banana 2 Lite: our fastest and cheapest Gemini Image model 🔘 Gemini Omni Flash: now available via the Gemini API and in @GoogleAIStudio to help developers generate and edit high-quality videos.

🟡 𝕏 @xai: Introducing Voice Agent Builder: a no-code platform to create human-like voice agents with Grok Voice. Available today at $0.05 / min. http://x.ai/voice — score 60 Sources: twitter_rss

Introducing Voice Agent Builder: a no-code platform to create human-like voice agents with Grok Voice. Available today at $0.05 / min. http://x.ai/voice

🟡 💬 New PyMuPDF release, supports Markdown [N] — score 56 Sources: reddit/r/MachineLearning

https://pymupdf.io/blog/markdown-in-pymupdf-1-28 PyMuPDF 1.28 release, introduces Markdown as a first class document in PyMuPDF. Seems useful for a variety of workflows. You can create PDFs from Markdown text with control over appearance using CSS

🟡 🏢 Redeploying Fable 5 Announcements Jun 30, 2026 Fable 5 returns globally July 1. We're also proposing an industry-wide framework for scoring jailbreak severity, together with Amazon, Microsoft, Google, and other Glasswing partners. — score 50 Sources: lab_blog/Anthropic

Product Jun 30, 2026 Introducing Claude Sonnet 5 Sonnet 5 delivers frontier performance across coding, agents, and professional work at scale. Announcements Jun 30, 2026 Claude Science, an AI workbench for scientists, is now available Claude Science is a customizable app that integrates the tools an

Omitted 2 additional model releases items from the main section; see raw data and source-specific sections below.

Developer Tools

🟡 💬 P Moth-Retrieval: Graph-Free Multi-Hop Retrieval via Query-Time Orchestration (Beating Graph-Based Systems on HotpotQA) [P] — score 69 Sources: reddit/r/MachineLearning

We just open-sourced MOTHRAG, a multi-hop RAG framework that skips the knowledge graph entirely. We kept hitting the same wall building multi-hop RAG: the systems with the best accuracy (GraphRAG, HippoRAG, RAPTOR) all lean on a knowledge graph built offline, and that’s great numbers, until the mome

🟡 🐙 HKUDS/VideoAgent — "VideoAgent: All-in-One Agentic Framework for Video Understanding, Editing, and Remaking" — score 66 Sources: github_trending

"VideoAgent: All-in-One Agentic Framework for Video Understanding, Editing, and Remaking"

🟡 💬 how many parameters can i run in my machine to make it as an ai agent assistant in my system — score 61 Sources: reddit/r/AIAgents

i wanted to do is to have an ai assistant that can do basic to moderate stuff, like to read whats on folders, create and structure them, maybe write some basic to intermediate code and commit and push it to github, maybe some shell commands. i am thinking of building another 2 or 3 agents to to act

🟡 💬 I extended Gemma4-31B to 44B (88 layers) — since Google won't give us anything bigger than 31B — score 54 Sources: reddit/r/LocalLLaMA

I've been just sit on this thread for a while now, both as a reader and occasional poster, so I figured it was finally time to share something I've been working on last weekends. Google hasn't shipped a dense Gemma4 bigger than 31B, so I decided to just build one myself. Heads up though — I'm not a

🟡 🐙 getmaxun/maxun — 🔥 The open-source no-code platform for web scraping, crawling, search and AI data extraction • Turn websites into structured APIs in minutes 🔥 — score 51 Sources: github_trending

🔥 The open-source no-code platform for web scraping, crawling, search and AI data extraction • Turn websites into structured APIs in minutes 🔥

Omitted 2 additional developer tools items from the main section; see raw data and source-specific sections below.

Business & Funding

🟡 💬 ACL ARR May 2026[D] — score 44 Sources: reddit/r/MachineLearning

Hi everyone. Do the ACL arr may 2026 reviews come out of July 2nd or do they come out on July 7 th?? How much does one need to get into Main or Findings? I am a bit new to this. Thanks a lot folks.

Research Papers

🟡 🤗 TRIAGE: Role-Typed Credit Assignment for Agentic Reinforcement Learning — score 68 Sources: huggingface · arxiv/cs.AI

Agentic reinforcement learning requires assigning credit to environment-facing actions such as searches, clicks, edits, navigation commands, and object interactions. Standard GRPO uses the final verifier outcome as a uniform advantage over all action tokens. This outcome signal is useful but structu

Other Signals

🟡 💬 Deepseek V4 Flash 2, 3 and 4 bits GGUFs — score 68 Sources: reddit/r/LocalLLaMA

🟡 𝕏 @AnthropicAI: Claude Fable 5 will be available again globally tomorrow. After a series of productive conversations with the US government, we're redeploying the model with a new set of classifiers to target and blo — score 50 Sources: twitter_rss

Claude Fable 5 will be available again globally tomorrow. After a series of productive conversations with the US government, we're redeploying the model with a new set of classifiers to target and block more cybersecurity tasks. In the near term, some routine tasks like coding and debugging will fal

🟢 Incremental

Model Releases

🟢 💬 End of an Agony. Real production service that uses LLM to earn money my team had made and now we are so happy that it will die. Here are some of my final "experiences". — score 39 Sources: reddit/r/LocalLLaMA

Hello everyone. I had posted in this sub about making a production service about 8 months ago. Here the link of my previous post . The idea was the same. We wanted to make a real production ser

🟢 💬 Looking for real experiences with AI chat agents in the NDIS space — score 39 Sources: reddit/r/AIAgents

Quick question for anyone working with an NDIS provider. We're seriously considering adding an ai chat agent this quarter because enquiries keep growing and our admin team is stretched. Nothing crazy, we just want people to get answers without waiting forever. I've narrowed it down to AusGPT and Exp

🟢 💬 gemma-4-31B on Cerebras is better than ChatGPT voice mode — score 29 Sources: reddit/r/LocalLLaMA

open models will win on inference too 🚀

🟢 💬 A system-level approach to prompt injection: separating instruction and data channels in LLM agents [P] — score 12 Sources: reddit/r/MachineLearning

Prompt injection has emerged as one of the most persistent failure modes in tool-using LLM systems, particularly in agentic workflows where models interact with external data sources. Most mitigation strategies focus on input filtering or model-side alignment, but these approaches struggle because t

Developer Tools

🟢 🐙 vas3k/TaxHacker — Self-hosted AI accounting app. LLM analyzer for receipts, invoices, transactions with custom prompts and categories — score 39 Sources: github_trending

Self-hosted AI accounting app. LLM analyzer for receipts, invoices, transactions with custom prompts and categories

🟢 💬 ZCode: New Agentic Code Editor from the Makers of GLM — score 29 Sources: reddit/r/LocalLLaMA

🟢 🐙 RedPlanetHQ/core — Your Personal AI OS — score 24 Sources: github_trending

Your Personal AI OS

🟢 💬 Ways to reduce token cost in AI agents — score 22 Sources: reddit/r/AIAgents

Ive been building AI agents for a while and noticed a few patterns that help cut token usage without wrecking the workflow. A few things that seem to matter most: - Set hard token budgets per task or step. - Stop runaway loops early with guardrails or circuit breakers. - Use smaller models for ch

🟢 🐙 togatoga/karukan — Japanese Input Method System for Linux, macOS, Neural Kana-Kanji Conversion Engine — score 21 Sources: github_trending

Japanese Input Method System for Linux, macOS, Neural Kana-Kanji Conversion Engine

Omitted 3 additional developer tools items from the main section; see raw data and source-specific sections below.

Infrastructure & Compute

🟢 💬 Anyone using TensTorrent gpus for your local ai? What's been your experience? — score 4 Sources: reddit/r/LocalLLaMA

I'm always keeping an eye on competitive hardware and was looking at tenstorrent cards, particularly the p150a which while its memory bandwidth is only 512GB/s, it does have 32 GB of GDDR6 and a high-speed Ethernet fabric (4×800 GbE) so multi-card systems don't rely on PCIe alone. something like thi

Research Papers

🟢 🤗 SpheRoPE: Zero-Shot Optimization-Free 360 Panorama Generation with Spherical RoPE — score 20 Sources: huggingface

We present a zero-shot, training-free and optimization-free framework for generating 360 panoramic images and videos by directly injecting spherical priors into pre-trained diffusion transformers. Existing methods either rely on costly fine-tuning on scarce panoramic data that limits generalization,

🟢 🤗 Goku: A Million-Scale Universal Dataset and Benchmark for Instruction-Based Video Editing — score 20 Sources: huggingface

Existing instruction-based video editing datasets commonly focus on single-task appearance editing, failing to meet the complex creative demands of real-world scenarios. To bridge this gap, we present Goku, a large-scale dataset featuring 2 million high-quality, instruction-aligned video editing pai

Other Signals

🟢 💬 ICML qr code visible [D] — score 31 Sources: reddit/r/MachineLearning

Hi everyone, The check in QR code is visible at my profile despite that my card isn’t accepting the payment transaction. What does that even mean? Thanks!

🟢 💬 Senior SWE Bench: a new benchmark focussed on realistically underspecified feature tasks — score 18 Sources: reddit/r/LocalLLaMA

🟢 💬 How to describe a model that has higher accuracy with fewer #param and FLOPs? [D] — score 12 Sources: reddit/r/MachineLearning

Hello, My supervisor is nowhere to be found so I am turning to the internet for my naive questions.

📈 Trending Repos

Repo	Description	Stars Today	Language
allenai/olmocr	Toolkit for linearizing PDFs for LLM datasets/training	295	python
virgiliojr94/book-to-skill	Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work.	205	python
open-webui/open-webui	User-friendly AI Interface (Supports Ollama, OpenAI API, ...)	171	python
HKUDS/VideoAgent	"VideoAgent: All-in-One Agentic Framework for Video Understanding, Editing, and Remaking"	150	python
getmaxun/maxun	🔥 The open-source no-code platform for web scraping, crawling, search and AI data extraction • Turn websites into structured APIs in minutes 🔥	88	typescript
vas3k/TaxHacker	Self-hosted AI accounting app. LLM analyzer for receipts, invoices, transactions with custom prompts and categories	60	typescript
RedPlanetHQ/core	Your Personal AI OS	32	typescript
togatoga/karukan	Japanese Input Method System for Linux, macOS, Neural Kana-Kanji Conversion Engine	29	rust

📄 New Papers

Title	Category	Hotness	Link
TRIAGE: Role-Typed Credit Assignment for Agentic Reinforcement Learning	research_paper	7	Open
What Drives Interactive Improvement from Feedback?	cs.AI	0	Open
Contrastive Reflection for Iterative Prompt Optimization	cs.AI	0	Open
How Can AI Find My Model? A Model-Finding Experimental Study Considering Data Formats, Embeddings, and Retrieval Strategies	cs.AI	0	Open
BayesBench: Evaluating LLM Belief Trajectories Under Multi-Turn Evidence Accumulation	cs.AI	0	Open
When Does Learning to Stop Help? A Cost-Aware Study of Early Exits in Reasoning Models	cs.AI	0	Open
Beyond expert users: agents should help users construct preferences, not just elicit them	cs.AI	0	Open
Investigating Multi-Agent Deliberation in Law	cs.AI	0	Open
Why Solve It Twice? Hierarchical Accumulation of Skills for Transfer-Efficient ML Engineering	cs.AI	0	Open
RoPoLL: Robust Panel of LLM Judges	cs.AI	0	Open
AgRefactor: Self-Evolving Agentic Workflow for HLS Compatibility and Performance	cs.AI	0	Open
Neuro-Bayesian-Symbolic Residual Attention Shallow Network: Explainable Deep Learning for Cybersecurity Risk Assessment	cs.AI	0	Open
HyPOLE: Hyperproperty-Guided Multi-Agent Reinforcement Learning under Partial Observation	cs.AI	0	Open
AgentBound: Verifiable Behavioral Governance for Autonomous AI Agents	cs.AI	0	Open
When Regulation Has Memory: Hysteresis and Control Burden in Artificial Agency	cs.AI	0	Open

🏢 Lab Blog Posts

• Anthropic: Redeploying Fable 5 Announcements Jun 30, 2026 Fable 5 returns globally July 1. We're also proposing an industry-wide framework for scoring jailbreak severity, together with Amazon, Microsoft, Google, and other Glasswing partners.

🐦 Twitter/X Highlights

Account	Tweet Summary
GoogleDeepMind	We’re shipping 2 major releases: 🔘 Nano Banana 2 Lite: our fastest and cheapest Gemini Image model 🔘 Gemini Omni Flash: now available via the Gemini API and in @GoogleAIStudio to help developers generate and edit high-quality videos. Post
xai	Introducing Voice Agent Builder: a no-code platform to create human-like voice agents with Grok Voice. Available today at $0.05 / min. http://x.ai/voice Post
AnthropicAI	Claude Fable 5 will be available again globally tomorrow. After a series of productive conversations with the US government, we're redeploying the model with a new set of classifiers to target and block more cybersecurity tasks. In the near term, some routine tasks like coding and debugging will fal Post
AnthropicAI	We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5. We'll begin restoring access tomorrow, and will share an update soon. We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models. Post
sama	Good new first: Sol is a smart, efficient, and a significant step forward. It is the same price as GPT-5.5. Also launching in the GPT-5.6 family is Terra, with 5.5-level performance at half the price. Bad news: at the request of the US government, it is launching today in limited preview instead of Post

Newsletter

• tldr: 12TB OF AI CODING AGENT LOGS (17 MINUTE VIDEO) - first seen 2026-06-30
• tldr: WE RAN 250 AI AGENT EVALS TO FIND OUT IF SKILLS BEAT DOCS. THE ANSWER IS MORE COMPLICATED THAN WE EXPECTED (6 MINUTE READ) - first seen 2026-06-30
• tldr: HOW WE USED DSPY TO TURN AI EVALUATIONS INTO BETTER RESPONSES IN DASH CHAT (5 MINUTE READ) - first seen 2026-06-30
• tldr: USING LOCAL CODING AGENTS (37 MINUTE READ) - first seen 2026-06-30
• tldr: TRUMP ADMINISTRATION ROLLS BACK PART OF ANTHROPIC MODEL BAN (3 MINUTE READ) - first seen 2026-06-30
• tldr: AI IS MAKING SILICON VALLEY PRODUCTIVE, ANXIOUS, AND AFRAID TO LOG OFF (11 MINUTE READ) - first seen 2026-06-30
• tldr: GOOGLE IS RATIONING GEMINI ACCESS TO META BECAUSE IT CANNOT PROVIDE ENOUGH COMPUTE (4 MINUTE READ) - first seen 2026-06-30
• tldr: AGENTICS/TECH THINGS: TOKENMAXXING IS DEAD, LONG LIVE TOKENMAXXING (18 MINUTE READ) - first seen 2026-06-30
• tldr: WHAT HAPPENED AFTER 2,000 PEOPLE TRIED TO HACK MY AI ASSISTANT (5 MINUTE READ) - first seen 2026-06-30
• tldr: SOFTWARE ENGINEERING IN THE AGE OF AI (15 MINUTE READ) - first seen 2026-06-30
• tldr: WHAT IT MEANS TO BE A MATHEMATICIAN WHEN AI DOES THE MATH (16 MINUTE READ) - first seen 2026-06-30
• tldr: ADOBE IS BUYING TOPAZ LABS, THE AI VIDEO ENHANCER (4 MINUTE READ) - first seen 2026-06-30
• tldr: FIGMA'S DESIGN AGENT, NOW WITH CUSTOM TOOLS AND GREATER CONTEXT (8 MINUTE READ) - first seen 2026-06-30
• tldr: THE LAYERS OF AI EXPERIENCE (21 MINUTE READ) - first seen 2026-06-30
• tldr: AI THUMBNAIL MAKER FOR YOUTUBE (WEBSITE) - first seen 2026-06-30
• tldr: AI-GENERATED VIDEO CREATION PLATFORM (WEBSITE) - first seen 2026-06-30
• tldr: OPEN SOURCE, APIS, AND THE RISE OF AGENT-LED GROWTH (11 MINUTE READ) - first seen 2026-06-30
• tldr: WE BOOKED 614 MEETINGS WITH ONE INBOUND AGENT. YOUR “CONTACT US” FORM IS COSTING YOU DEALS (11 MINUTE READ) - first seen 2026-06-30
• tldr: THE AI INDUSTRY AS YOU KNOW IT DIED TODAY (7 MINUTE READ) - first seen 2026-06-30
• tldr: AI SHOULDN'T SHRINK HEADCOUNT. IT SHOULD SHRINK TEAMS (5 MINUTE READ) - first seen 2026-06-30
• tldr: US GOVERNMENT ALLOWS ANTHROPIC LIMITED RELEASE OF AI MODEL THAT SPARKED CYBERSECURITY CONCERNS (3 MINUTE READ) - first seen 2026-06-30
• tldr: ENTERPRISE AI SPENDING IS STILL HEATING UP (4 MINUTE READ) - first seen 2026-06-30
• tldr: HP EXPANDS OPENAI FRONTIER PARTNERSHIP (3 MINUTE READ) - first seen 2026-06-30
• tldr: GOOGLE ANTIGRAVITY AGENTS GET FULL CONTEXT WITH GITLAB ORBIT (5 MINUTE READ) - first seen 2026-06-30
• tldr: OKTA BRINGS AI AGENT GOVERNANCE TO FEDRAMP AND HIPAA CUSTOMERS (3 MINUTE READ) - first seen 2026-06-30
• rundown-ai: OpenAI claims to have trained protections into 5.6, but METR did find some gaps, including Sol cheating its evals at a rate higher than any other model. - first seen 2026-06-29
• rundown-ai: AI-powered movie production with Claude - first seen 2026-06-29
• rundown-ai: 1. Download and install Palmier Pro, open a new project, and create an Anthropic API key at console.anthropic.com/settings/keys - first seen 2026-06-29
• rundown-ai: Build production agents in minutes - first seen 2026-06-29
• rundown-ai: AI economy banked $110B last year - first seen 2026-06-29
• rundown-ai: The Rundown: New Exponential View research, analyzing data from various sources, shows that the generative AI industry hit $110B in revenues last year and is on track to touch $175B, scaling 3x faster than any prior tech - first seen 2026-06-29
• rundown-ai: MAI-Code-1-Flash - Microsoft’s in-house coding AI, GA for select users - first seen 2026-06-29
• rundown-ai: Elon Musk said Grok 4.5, trained with supplemental Cursor data, is in private beta at SpaceX and Tesla, claiming its performance matches Anthropic’s Claude Opus. - first seen 2026-06-30
• rundown-ai: Google reportedly capped Meta’s Gemini usage as soaring demand for AI compute outpaced available capacity, delaying some of Meta’s internal projects. - first seen 2026-06-30
• rundown-ai: OpenAI reset usage limits for all Codex users after mitigating a fraud detection bug that caused some accounts to burn through quotas faster than intended. - first seen 2026-06-30
• rundown-ai: Austria proposed hosting Anthropic in the EU after U.S. curbs on Fable 5 and Mythos 5, arguing Europe needs independent access to frontier AI. - first seen 2026-06-30
• rundown-ai: Read our last AI newsletter: White House reins in OpenAI’s GPT-5.6 - first seen 2026-06-30
• rundown-ai: RSVP to next workshop on June 30: Master AI video editing - first seen 2026-06-29
• rundown-ai: OpenAI's most powerful model is here — but not for everyone - first seen 2026-06-30
• rundown-ai: Cursor takes coding on the move with new iOS app - first seen 2026-06-30
• rundown-ai: The move trails mobile launches by Anthropic and OpenAI, with Claude Code lead Boris Cherny recently saying, "Most of my coding now is on my phone." - first seen 2026-06-30
• rundown-ai: 1. Install Codex, sign in, and pick a task like uploading a YT video or exporting reports. Then, open Plugins, install Record & Replay and Computer Use - first seen 2026-06-30
• rundown-ai: Kling brings cinematic AI production to Cannes - first seen 2026-06-30
• rundown-ai: Anthropic clocks Claude's daily grind - first seen 2026-06-30
• rundown-ai: OneTrust AI Governance - Turn AI risk into enforceable controls. Named a Visionary by Gartner - first seen 2026-06-30
• rundown-ai: Cursor for iOS - Cursor's new mobile experience for coding on-the-go - first seen 2026-06-30
• rundown-ai: Cognition launched Devin Fusion in preview, a new coding harness that pairs a frontier model with a cheaper "sidekick" agent to maintain quality at lower cost. - first seen 2026-06-30
• rundown-ai: Apple's Vision Pro chief Paul Meade is leaving Apple for OpenAI's hardware unit, reuniting with ex-colleagues Jony Ive and Tang Tan to build its AI-powered devices. - first seen 2026-06-30

Repeated From Recent Briefings

• usestrix/strix — Open-source AI penetration testing tool to find and fix your app’s vulnerabilities. - first seen 2026-06-28
• diegosouzapw/OmniRoute — Never stop coding. Free AI gateway: one endpoint, 231+ providers (50+ free), connect Claude Code, Codex, Cursor, Cline & Copilot to FREE Claude/GPT/Gemini. RTK+Caveman stacked compression saves 15-95% tokens, smart auto-fallback, MCP/A2A, multimodal APIs, Desktop/PWA. - first seen 2026-05-08
• Does VLA Even Know the Basics? Measuring Commonsense and World Knowledge Retention in Vision-Language-Action Models - first seen 2026-06-18
• facebook/astryx — An open source design system that's fully customizable and agent ready - first seen 2026-06-27
• browser-use/video-use — Edit videos with coding agents - first seen 2026-06-27
• HKUDS/Vibe-Trading — "Vibe-Trading: Your Personal Trading Agent" - first seen 2026-06-03
• google/agents-cli — The CLI and skills that turn any coding assistant into an expert at creating, evaluating, and deploying AI agents on Google Cloud. - first seen 2026-06-30
• farion1231/cc-switch — A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Gemini CLI & Hermes Agent. Only official website: ccswitch.io - first seen 2026-05-08
• Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning - first seen 2026-06-30
• ogulcancelik/herdr — agent multiplexer that lives in your terminal. - first seen 2026-06-21
• ... plus 172 more repeated items in processed data