Tools Bench.

Product launches and open-source repos with enough signal to earn a second look.

Last Brew Time: Aug 27, 2026, 11:13 AM PT

Insight

Builders this run are quietly turning expertise itself into an installable file format

01.PACKAGING KNOW-HOW AS A SKILL

02.TEACHING AGENTS TO COORDINATE INSTEAD OF COLLIDE

03.WRAPPING SOMEONE ELSE'S OPEN MODEL IN A FRESH UI

Featured

GitHub92.2K

thedotmack/claude-mem

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

github
Persistent Context Across Sessions for Every Agent

Market Signal

Why It Has Market Pull

claude-mem is a genuinely high-momentum, actively-shipping open-source project with real multi-platform buzz (X, YouTube, Medium) and a large, engaged user base — this is a legitimate one to track closely, not a star-count mirage.

  • 92,197 stars and 8,105 forks as of 2026-08-27 — a fork ratio (~8.8%) consistent with organic large-scale usage
  • 172 contributors with a real graduated commit distribution, not a single-author shell repo
  • 10 releases in the past ~5 weeks, most recent (v13.16.1) shipped 2026-08-26, one day before this review
  • 312 open GitHub issues, many containing detailed real-world bug reports
  • Multiple independent YouTube tutorials plus a Medium tutorial and a dedicated X account (@Claude_Memory)

feedbacks

What People Are Saying

  • "You can now give infinite memory to Claude Code. Claude-Mem just released a free open source memory plugin"X post

  • "Multiple errors and issues encountered during install, uninstall, reinstall"GitHub issue

  • "244 consecutive failures observed; zero memories saved for ~90 minutes"GitHub issue

  • "882 lost observations over six days documented"GitHub issue

  • "Codex SessionStart hook silently disables all capture"GitHub issue

  • "Install aborts on npm 12 with EALLOWSCRIPTS, leaving worker stopped"GitHub issue

  • "This Plugin Gives Claude UNLIMITED Memory!"YouTube video title

GitHub34.6K

anthropics/claude-plugins-official

Official, Anthropic-managed directory of high quality Claude Code Plugins.

Market Signal

Why It Has Market Pull

This is Anthropic's own first-party plugin directory, pre-installed by default in Claude Code, and it shows real day-to-day usage rather than a one-off launch bump. Activity is dense and current, with organic outside discussion and a parallel ecosystem of community-built marketplaces growing around it. This looks like durable core infrastructure, not a hype repo.

  • 34,623 stars and 3,908 forks, with 1,039 open issues and 102 open pull requests - issue volume this high points to genuine day-to-day use, not just star collection
  • Commit history shows near-daily activity through August 27, 2026, mixing automated dependency bumps with human-reviewed partner-plugin PRs
  • Confirmed genuinely official: announced by Anthropic on October 9, 2025 and auto-added as the default marketplace the first time Claude Code runs interactively
  • Sparked a standalone Hacker News thread plus independent community marketplace sites built specifically to complement it
  • A separate, more open community repo exists for third-party submissions that pass automated safety screening

feedbacks

What People Are Saying

  • "[Bug] GitHub plugin fails: 'Incompatible auth server: does not support dynamic client registration'"GitHub issue

  • "All LSP plugins missing .lsp.json — installed plugins have no LSP configuration"GitHub issue

  • "GitHub plugin missing README with authentication setup instructions"GitHub issue

  • "jdtls-lsp generates Eclipse metadata files in project root"GitHub issue

  • "plugin maintainers do not want to have to build a marketplace as well as a plugin"HN comment

  • "using the full HTTPS URL worked instead for adding the marketplace"HN comment

  • "Anthropic does not control what MCP servers, files, or other software are included in plugins and cannot verify that they will work as intended"GitHub repo disclaimer

Product Hunt338

PostHog Desktop

A multiplayer workspace for you, your team, and your agents – with your product data as context, and PostHog tools to ship and measure. Build and edit your product | Run a fleet of agents | Turn product signals into PRs.

Market Signal

Why It Has Market Pull

PostHog Desktop is a new multiplayer, agent-driven workspace from PostHog, the well-funded, Y Combinator-backed open-source product analytics company. Among similar launches, it stands out for an unusually credible foundation, even though the desktop app itself is still early and rough around the edges.

  • 338 upvotes and #3 Day Rank on Product Hunt (August 26, 2026)
  • Parent company PostHog has raised $194M total, including a $75M Series E at a $1.4B valuation (September 2025)
  • Main posthog/posthog GitHub repository has roughly 38,800 stars
  • Desktop's dedicated repo logged 2,489 commits and 65 forks before being folded into the main monorepo
  • Y Combinator Winter 2020 alum with an established customer base across analytics, session replay, and feature flags

feedbacks

What People Are Saying

  • "Great idea, congrats on the launch!"Product Hunt comment

  • "Good one"Product Hunt comment

  • "used PostHog to build Veltrix AI ... replaced three separate tools"Product Hunt review

  • "an AI-powered product editor for product builders"Product Hunt comment

  • "pre-alpha and not production-ready"GitHub repo description

  • "This repository has migrated to the PostHog monorepo"GitHub archive notice

Sources

GitHub

DietrichGebert/ponytail

113.8K

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

K-Dense-AI/scientific-agent-skills

35.2K

Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 175,000+ scientists worldwide. 163 ready-to-use validated skills plus 100+ scientific databases covering biology, chemistry, medicine, and drug discovery. Compatible with Cursor, Claude Code, Codex, Pi, Antigravity, and the open Agent Skills standard.

freestylefly/awesome-gpt-image-2

22.9K

Prompt as Code | GPT-Image2 工业级提示词引擎与模板库，530+ 个案例逆向工程，20+ 套工业级模板，并提炼出Skills，持续更新中

Product Hunt

x1

491

X1 is an AI app builder that takes you from an idea to an iPhone app you can publish on the App Store. Instead of trying to generate the entire app from one prompt, X1 guides you step by step. It asks focused questions, designs each screen for you to review and edit, then builds the app in stages. Preview and test it on your iPhone as you go. When it’s ready, X1 prepares your App Store listing, screenshots, and submission. No coding required.

Expertise AI

362

Expertise AI is where GTM experts monetize their playbooks as protected Al skills. Each expert gets a storefront where businesses demo and install these skills on a subscription, and gets paid every time one runs.

ChatCut Desktop

288

Introducing ChatCut Desktop, the AI video editor built for humans and agents to edit together. Use ChatCut’s built-in agent, or connect ChatGPT/Codex or Claude Code. Say how you want your video edited and watch them happen on a fully editable timeline. Edit footage, create motion graphics and captions, generate video, images, music and SFX, and save reusable editing skills. Everything runs locally, and you stay in control. Export XML to Premiere Pro, DaVinci Resolve or CapCut.

OpenComputer

138

Introducing OpenComputer Deploy your agent as a function. Get a computer for it. - a real Linux machine per session - ffmpeg, chromium, git, install anything - durable: hibernate + resume - your keys never enter the runtime The cloud of the last decade was built for apps. The next one is for trillions of always on agents. OpenComputer is their home.

ify

116

Most AI support tools ask you to rip out your helpdesk first. Ify works on top of Freshdesk, Zendesk, Salesforce, or HubSpot and gets to work across email, chat, WhatsApp, Slack, and more. The part that usually stalls an AI support rollout is "our docs aren't good enough." Ify solves that by building its own knowledge base: it scrapes your site and docs, generates SOPs from release notes and past ticket resolutions. Built for support teams who want something live now.

Evidence Core

113

Evidence Core is an open source framework for building business intelligence applications with coding agents. Build your entire business intelligence platform as a repo of files, and deploy it anywhere, for free. Metrics, dashboards, reports and themes are all specified in code which you can validate, preview and serve using the Evidence CLI.

YC Launch

Denta 🦷 Dental Insurance

4

Save time and money on dental insurance Denta · Summer 2026 · Fintech Tags: Fintech, Health Insurance. Website: https://denta.com

Hacker News

Keenable – A different web search API for AI agents

12 pts

Hey HN! We built https://keenable.ai , a different web search API for AI agents. Keenable searches our own 100B+ page index. We are focused on low cost and latency (p95 <250ms from us-east). We don’t believe in benchmaxxing, so we open-sourced our internal benchmarking suite, NEEDLE (available at https://keenableai.github.io/needle ): a live benchmark that compares Keenable with other search APIs on fresh agent-like queries. I spent seven years at Amazon as a scientist working on web grounding f... (12 points, 5 comments).

Sparrow-2 – Solving the cocktail party problem

6 pts

Hey there, I’m Brian. I've been shipping conversational models over here at Tavus for the past two years. I want to tell you about our new audio-understanding model: Sparrow-2! It’s a new category of model and a unique new approach to conversational audio. Earlier this year we launched Sparrow-1, (at the time) our SoTA turn taking model. Since our Sparrow-1 launch, I’ve spent a lot of time listening to humans talking and trying to really understand how people know when to talk, when to listen, a... (6 points, 0 comments).

OzBrain, a shared brain for knowledge between agents and your team

92 pts

I think agent-first chat interfaces will be a primary software modality and busy dashboard/UI will go away. I’m not sure who exactly wins it, but I want my knowledge to grow/go with me. A lot of the “knowledge” ie research, analysis, reasoning will be done by agents as the primary user. Our current notes tools & tasks management systems were built for humans… I don’t care what the 17th thing on my bug backlog is. I want to conduct agents that can execute for me and do great work. What I built Oz... (92 points, 55 comments).

I shrank DeepSeek V4 Flash to 57GB and it wrote a compiler on my Mac

21 pts

I built a specialized package of DeepSeek V4 Flash 0731 (originally 284B total parameters, 13B active), preserving reasoning, tool calling and coding capabilities: https://huggingface.co/steadfastgaze/DeepSeek-V4-Flash-0731-... I let it write a minimal C compiler targeting ARM64, then test the result with Fibonacci and FizzBuzz programs, and it succeeded in less than 1 hour, with the full recording at: https://youtu.be/XiwSilmV8B0 You can run it on Silicon Macs with my engine https://github.com/... (21 points, 3 comments).

LilScript makes JavaScript libraries smaller

14 pts

https://yeargun.github.io/lilscript/ LilScript is a typed, compression-first language that compiles into js and sometimes into exec(will be more stable in future). The compiler mangles, reshapes the program into optimized js that happens to be 5-15% smaller (after gzip/br compression or raw) compared to the best performing JS toolchains like oxc/esbuild/terser/.. ## What has been proven to work with LilScript? - makes VSCode's core js modules 20% smaller on average - makes the world's most perfo... (14 points, 2 comments).

Concord – let Claude Code, Codex and Cursor talk to each other

9 pts

Recently I've been running more and more agents in parallel however I noticed that they have no task context of what the other agents are doing even when a lot of work is interconnected It's like taking Slack away from a team. Agents duplicate work, make conflicting changes, and step on each others' toes simply because they can't talk to each other. Concord is an MCP + CLI that lets coding agents claim work, see what other agents are doing, and message each other live. (9 points, 3 comments).

HF Spaces

Qwen-Image-Edit-2511-LoRAs-Fast

2.7K likes

Demo of the Collection of Qwen Image Edit LoRAs Qwen-Image-Edit-2511-LoRAs-Fast is a Hugging Face Space tagged with gradio, mcp-server, region:us. It has 2674 likes on Hugging Face.

Agent Memory Leaderboard

767 likes

Unified memory evaluation · Results expected August 12. Agent Memory Leaderboard · 记忆之巅 A unified, open, and reproducible evaluation platform for long-term memory systems and memory-enabled agents. Agent Memory Leaderboard (AML) compares research methods and commercial products under one evaluation contract. Candidate systems implement memory Add and Search; the official platform fixes Answer, Eval, datasets, models, configurations, result review, and publication. > First public release: The inaugural verified leaderboard is expected to be published on August 12, 2026. > 首期发布： 首期经核验榜单预计将于 2026 年 8 月 12 日发布。 Results are separated along two independent dimensions. Textual and coding tasks use...

MiniMax H3 Turbo LoRA

292 likes

Video generation with a synchronized soundtrack MiniMax-H3 — unquantized, split across two Spaces Joint video and soundtrack out of a single denoising pass, at bfloat16 with no quantization anywhere. This Space is the denoising half: the 61.73 GiB transformer and the two autoencoders. The 62.14 GiB Qwen3-VL conditioner runs in qwen3vl-conditioner, which this Space calls over the gradio API for every request. The weights are the public MiniMaxAI/MiniMax-H3 diffusers checkpoint. MiniMax-H3 is 195.9 GiB in bfloat16 and a ZeroGPU Space is evicted at 150 GB of storage. An unquantized single Space is therefore impossible, which is why quantized demos of it run NVFP4 or float8 weights. Cut the Mini...

MiniMax Music 3 Studio

285 likes

MiniMax Music 3 Studio — diffusers demo Streams full songs from lyrics + a structured caption using the MiniMaxMusic3Pipeline diffusers port. The input surface is a single Suno-inspired custom gr.HTML composer (Simple ↔ Studio modes, section-tag chips, structured-caption fields per the official prompting guide) that drives Gradio events via trigger()/props.value; styling uses only theme CSS vars so it follows the Citrus theme natively. Weights: MiniMaxAI/MiniMax-Music3 AoTI kernels: diffusers-internal-dev/MiniMax-Music3-aoti (compiled on RTX Pro 6000, matching ZeroGPU hardware) Generation streams chunk by chunk with a configurable playback headroom. The 8B language-model stage runs eager on....

SenseNova-U1.5-8B-MoT

60 likes

Unified text-to-image and image editing model Text-to-image and image editing demo for sensenova/SenseNova-U1.5-8B-MoT, a natively unified multimodal model (18B params, bf16) built on the NEO-unify architecture. Leave the image upload empty for text-to-image generation, or upload one or more images and write an edit instruction for image editing. Advanced options exposes denoising steps, guidance scale, timestep shift, image guidance (editing) and the seed. The step count defaults to 28 rather than the model card's 50: a fixed-seed A/B found 28 keeps composition, prompt adherence and text rendering intact — losing only some micro-texture in landscape and skin, and nothing measurable when edi...

Microduck Sandbox

48 likes

Real trained RL policies for the Microduck robot, running fully in the browser: MuJoCo compiled to WebAssembly steps the physics, onnxruntime-web runs the policy network at 50 Hz. No server, no backend. Two locomotion variants of the same robot are included: legs (walking, the default) and rollers (the wheeled skating variant). Press M (or hold D-pad up ~1 s on a gamepad) to switch; the roller model, meshes and policies are lazy-loaded on the first switch. | Mode | Checkpoint | What it does | |--------|-----------|--------------| | Run (legs) | BESTalphawalking.onnx | Velocity-tracking locomotion (arrows / WASD to steer) | | Sit | BESTalphasitstand.onnx | Sits down on its hull, stands back u...