tierlist.tech
中EN

#AI 智能体

200

2026-09-06 okf-memory/okf-agent-memory S Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.
2026-10-22 Show HN: Omnara – Open-source alternative to Claude Managed Agents S Hey all, Kartik from Omnara ( https://omnara.com) here. We're building an open-source API for running durable agents. You create an agent with an API call, and Omnara runs the agent loop, gives the agent sandboxes to run commands in, and handles secrets, permissions, and RBAC. Each agent's state is stored in Postgres, so agents survive restarts and cost nothing while idle. You can message them and stream their events over the API. You may remember us from our early days, when we made a remote-control phone app for Claude Code and Codex. Since the harness ran on a laptop back then, we had to build a relay to reach it, it stopped when the laptop closed, and moving it to a cloud sandbox meant syncing code and agent state between machines. We didn't think remote-controlling a local agent was end-game for how people would run agents, so we built a new product where the agent runs in the cloud, any client can talk to it through an API, and it can still use your laptop as one of its machines. Anthropic and OpenAI are building hosted versions of this, like Claude Managed Agents and the OpenAI Agents API. Omnara is an open-source version you can run on your own servers and use your own models. It's basically a harness server like the opencode or codex servers, but durable, scalable, with the usual harness features like file and shell tools, subagents, compaction, permission prompts, MCP, skills, memory, yada yada yada. Here's some interesting architectural decisions we made while building this out: Most harnesses, like Claude Code, Codex, OpenCode, Pi, Goose, etc, run the agent loop in one long-lived process that keeps its log in a JSONL or SQLite file. In Omnara, the log is an append-only event log in Postgres, with inputs, model outputs and tool calls stored as typed rows in their own tables. The loop runs on stateless workers in the control plane. A worker runs one iteration of the loop, either a model call or the tool calls it produced, then releases the agent and puts the next iteration on a queue in Postgres. Whichever worker is free picks it up, so each iteration of an agent's loop can run in a different process on a different machine. Each iteration's result commits in one transaction, so the model output, its tool calls, usage and the next wakeup are saved together or not at all. If a worker dies mid-iteration, nothing commits, its lease expires, and another worker retries it. Agents waiting on a user, a tool or a retry don't hold a worker, so one worker can multiplex thousands of agents. Also unlike other harnesses, the machines that the workers (our agent loop) run on aren't where the agent runs its commands. Instead, any macOS or Linux machine with an internet connection, like your laptop or any cloud sandbox, can run a small daemon (omnarad) that connects out to Omnara, so it doesn't need any open inbound ports. The agent then gets tools for that machine, like running a command or editing a file. An agent can have any number of machines connected to it. A fun thing I like to do is I have my laptop and a cloud sandbox connected to the same agent, so when my laptop is on, my agent codes there and I can interact with it, and I have it often push to Github. When I close my laptop, my agent sees the laptop went offline, pulls the latest changes to its cloud sandbox, and continues there. The entire codebase is open source at https://github.com/omnara-ai/omnara, and it’s self hostable, even on a single machine with docker compose. You can also use the hosted version at https://omnara.com. Let me know what you think!
2026-09-03 anthropics/commerce-agents S Reference blueprint for building shopping and merchant agents with Claude. Examples in retail, commerce, telecom, and entertainment included.
2026-10-01 feder-cr/dots S Open-source dots for the web: an AI agent with its own browser, one that does not get blocked.
2026-09-20 Show HN: CUA-S1 – A System One Model for Computer Use S Hello HN! We're Dillon and Francesco from Cua. We were wondering how many computer use tasks actually need a full general purpose LLM (e.g. gpt-6-astra, claude-opus-5 etc.) to think through all their decisions and steps. Some tasks require thinking about a plan, exploring different paths, recovering from failure. Other tasks are a question of making local decisions, like this value should go in this box, or should I check this box, or this element should be ignored. We wondered how far we could go with a small model trained to only make these kinds of decisions. Our inspiration was Typesafe's Jev and its System One Model framing. This is a nod to the dichotomy between thinking quickly, automatically, and intuitively (system 1) vs. thinking slowly, analytically (system 2), as described by Daniel Kahneman. The interesting question for us was: what happens if you give a model an interface of current context, and a set of possible choices, and you ask it to return a probability for each choice? This kind of model does not generate output token by token like most LLMs do, but rather scores the options you give it, which you can check, trust, and use to drive your app's behavior. CUA-S1 is our answer for narrow, specialized decision models for computer use. Our first release is CUA-S1-FORMS. We built this from ideas and code in jevlike, and then trained a second model just to handle form interactions. It has 706k parameters, and the original checkpoint is 2.8 MB. The first training iteration took less than 30 minutes on synthetic data. Given a set of structured elements and values extracted from a document, it predicts whether to use the given value, CHECK, CLICK, or SKIP for each element. It does not predict new values for text fields, and does not consider screenshots. Element decisions are scored together, and your code can order the actions, and Cua Driver will execute them one at a time. A first evaluation of this specialist vs. hosted Jev on our form task: - For the whole decision set: 99.7% correct vs 83.6%. - For the subset of steps that require an action: 100% correct vs 96%. - For the subset of steps that are just leaving already-filled fields alone: 100% correct vs 74%. The specialist was trained specifically for this task and convention (just press skip for already filled boxes), while hosted Jev has not been fine-tuned for it, so this is an experiment in scoped specialization. We measured 7-9 ms to score a form locally vs. 260-280 ms per call to hosted Jev including network latency, though those samples measure different things and are not end-to-end form completion times. Our interest here is in the space between a brittle script and a general agent loop. The content and layout of form fields vary enough that scripts get unwieldy, but the set of available decisions can remain narrow and well scoped. We want to explore the possibility of a general agent encountering something novel, and passing well understood decisions over to specialists like this. That is a direction we are looking into. The current release is for forms only. We're open sourced the synthetic data generation, training, evaluation, and Driver integration under libs/cua-s1 with an MIT license. Comments welcome! Especially if you are building computer-use agents and have run into a recurring decision that is too variable to script but is too narrow to call another LLM for.
2026-09-27 dzhng/jevgrep S Find code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.
2026-09-25 Show HN: Whiteboard (YC W26) – An open-source IDE for thoughtful software design S Hello! We’re Sid, Alex, Ketan, and Milan. We’re building Whiteboard ( https://whiteboard.dev.fast/), an open-source desktop app where humans and agents can architect software together in a common workspace. Here’s our repo: https://github.com/devdotfast/whiteboard. We were missing the feeling of a “whiteboard session” with another dev where you leave with a deep understanding of a system, so we built this app for ourselves. Whiteboard plugs into the tools you already use - e.g. Claude Code, Codex, etc. – and gives your agent an SDK to draw on an in-app canvas to describe its work. We began with an MVP based on HTML artifacts and started rethinking the app as we ran into limitations: 1. Built on top of CodeOSS: We found that in pure HTML tools it was hard to connect a spec or diagram to code. In Whiteboard, when you click on visualizations like a sequence diagram, an entity relationship diagram, or a quote from the agent’s trace, you can jump to the underlying code directly. When navigating code, you get keybindings and LSP support from VSCode out of the box. We’ve found this is especially valuable because tradeoffs are often only discovered after a first pass at implementation (re: slop) 2. Semantic diff viewer: we wrote a semantic, AST-aware diff viewer in Rust so you can only view the code changes which are relevant to you [1]. We’ve set up some sane defaults: large added functions are summarized as pseudocode, and things like unit tests and large documentation changes are collapsed / hidden. This is all customizable with a WASM-based plugin system. 3. Decision Log: We found it difficult to reason about what set of decisions our agents made autonomously. So we built tools for agents to query and link their own traces to the Whiteboard, so you can understand how the requirements that you set were implemented, and understand what decisions your agent made autonomously. Here’s a quick demo video explaining more: https://www.youtube.com/watch?v=ChPn3ftULWE Folks at companies like Salesforce and Modal are using Whiteboard today as a review tool for architecture or spec-level changes – really any change where they want to be involved: 1. Reviewing your own coding agent’s work: because Whiteboard makes it easier to review large amounts of code, folks will typically have their AI agents create a prototype and a corresponding Whiteboard session so they can iterate on the design. 2. Reviewing other people’s changes: We’ve found that Whiteboard is particularly helpful when composed with tools like Greptile. For example, you can run an automated code reviewer on small changes and escalate to a Whiteboard session for the changes that require human judgement. Why we built this: we’re four buddies from college who quit our jobs as tech leads right before agentic coding became industry standard. As we iterated towards an MVP for a previous idea, we struggled to maintain a comprehensible codebase while reaping all the velocity benefits of agentic coding. As more PRs were merged without our understanding, we felt a ‘cognitive debt’ begin to seep in, until it became difficult for us to even contribute to the system [2]. We’re releasing our desktop app under an MIT license. Please poke through and feel free to contribute! Eventually we’ll charge companies for a hosted web version that manages whiteboard session creation alongside features like trajectory storage and multiplayer reviews. Everything will always remain self-hostable. Thanks for reading, and we hope you try it out! We would love to hear any feedback and to learn from your expertise. Here’s are the project links again: https://github.com/devdotfast/whiteboard, and you can install (for MacOS + Linux) at https://install.dev.fast [1] diffs library: https://github.com/devdotfast/diffr [2] Credit for the term ‘cognitive debt’ goes to https://www.geoffreylitt.com/2026/07/02/understanding-is-the...
2026-10-13 onehumanbeing/esp32-iphone-agent S Let an AI model use a real iPhone with a $7 ESP32: AirPlay eyes, USB HID hands, any OpenAI-compatible vision model. No Mac, no developer account, no jailbreak.
2026-09-30 openai/mcp-extensions S Build plugins that feel like native, first-class features of ChatGPT.
2026-09-10 Show HN: Geiger – See every AI agent on your machine and what it can touch S
2026-10-17 feder-cr/git-ai S Git designed for AI agents.
2026-09-22 Show HN: Foremerge – Catch intent conflicts between parallel coding agents S At, GPTree, we run several coding agents across our team on one repo using parallel worktrees. Apart from wasted time reviewing and fixing conflicts at PR time, the failures that hurt the most are when multiple plans or tickets cause architecture changes that cannot both be true. Ex. one agent replaces a class while another one is in the process of extending it. Git only notices if the resulting patches happen to touch the same lines and the review only catches it if they are familiar with both tickets. Foremerge is a local "git like" coordination layer that sits above git (ie. does not interact with or change the way git and worktrees function), Before editing each agent publishes an intent and the scopes it will change, with the operation it plans to complete on each one. foremerge intent publish --agent "$A" \ --summary "Replace PaymentService with StripePaymentService" \ --scope symbol:PaymentService=replace foremerge intent publish --agent "$B" \ --summary "Add PayPal support to PaymentService" \ --scope symbol:PaymentService=extend The publish by the 2nd agent returns a HIGH destructive_vs_additive finding before writing any code. Agents keep their own worktrees and the shared state is one SQLite file in gits common direectory. No hooks, no merge drivers, nothing rewrites your history. It ships as one Rust binary with a CLI and MCP server with 18 tools and `foremerge setup all` wires it into Claude Code, Codex and Cursor. Because the protocol has nothing provider specific, a Claude agent and a Codex agent coordinate through the same store. Before any work is accepted, Foremerge runs a named check that you configured against the exact git state of the change. An agent that says tests pass is recorded but it dosnt satisfy the acceptance gate without running the check itself. Detection is deterministic, no judge model reading your code. HIGH conflicts are only asserted for declared operations, ie. matches inferred from prose cap out below high. Claims are advisory leases, not locks so two agents can still hold the same scope without deadlock. The open source version is single matching and so not a distributed consensus. We have tested this up to 98 parallel agents all working on the same repo with zero conflicts (was supposed to be 100 but 2 agents failed to run due to resource limitations) I replayed 76 intents on my own agents from a build last week in the order they happened. The sample had exactly 1 conflict (which was flagged) and the review found a blind spot where one agent claimed scope by class name and the other claimed it by an internal method. We are working on fixing that for the next release. Setup is a 30s install by pasting the quickstart instructions from the readme.md into your agent or manually: `curl -fsSL https://foremerge.com/install.sh | sh` or `cargo install --locked foremerge`, then `foremerge init && foremerge setup all` in a repo. Apache-2.0. The feedback I want most is which conflicts between your agents plans would you actually want flagged and which would you tollerate as noise? Repo here: https://github.com/naw103/foremerge Website: https://foremerge.com More information on the problems this solves: https://foremerge.com/blog/
2026-10-11 aindeev/agent-chrome-relay S Let your AI agents use your Chrome, signed in as you, in background tabs. By Alexey Indeev · blog: read.aindeev.com · X: @AlexeyIndeev
2026-10-06 elstongun/leviathan S **Deep memory for agents over large datasets.** Leviathan is a single static binary that turns your records (JSONL, JSON, CSV/TSV, SQLite, or anything a database CLI can export) into a ranked full-text index.
2026-10-01 Show HN: Open-source model routing for coding agents at Astra-level performance S A few months ago we started building a model router for coding agents because we thought we could outperform any single model with an ensemble approach. Recently we’ve achieved that milestone and I want to talk about how we did it. First of all, a quick explanation: the Weave Router ( https://github.com/weave-os/router) plugs into any coding agent (e.g. Claude Code or Codex) and intelligently switches between LLMs. So, for example, Astra handles tricky debugging or complex system design tasks, and Deepseek v4 Flash handles simple frontend updates. What we’re announcing today is our new routing model, which we’re calling Weave Router 2.0. We benchmarked 2.0 against GPT-6 Astra on Terminal Bench 4.0 and SWE Atlas. On both benchmarks, the router had equivalent pass rates. On Terminal Bench, the router hit 52% of Astra’s cost, and completed tasks 2.2x faster. On SWE Atlas, the router cost 54% as much as Astra and ran 2.5x faster. (Full results on our website at https://weaveos.com/router!) It turns out training a model to route effectively - taking into consideration model capabilities, costs, cache awareness, and more - is a really hard problem! I want to talk about three ways we were able to improve so much over the last few months: 1) a new architecture, 2) larger training data set size, and 3) smarter cache-eviction impact calculation. 1) a new architecture. Our initial approach used an RL model without many priors. While RL is still an important part of the story, the cost of fully exploring the space of routing decisions is very high, so we’ve taken some shortcuts that have significantly improved performance. Consider how large the search space for the routing problem is. Take a typical coding agent session, with ~100 agent turns (i.e. 100 LLM API calls). Technically there are 100 chances to select a model. If we assume a roster of ~10 models (of course there are lots more but we can remove any that are Pareto dominated), then there are 10^100 possible paths through that session. We simply cannot explore all of them! So that's why clever tricks to shrink this space are so important. In particular: we trained a hidden Markov model to trace the session state, then a classifier maps the session to one of a few buckets of similar models. Using the HMM allows us to evaluate not just where a session is currently, but how it got there. We've gotten significantly better performance on bucket selection by incorporating that information - we believe this is because two sessions that might look quite similar to a naive classifier are much better distinguished by this HMM approach. Using this HMM + classifier to select a bucket first significantly shrinks the space to explore, by throwing out most models that could not reasonably serve the given session. This rearchitecture was the single biggest performance unlock! 2) larger training data set size (much less technically interesting but still an important part of the story). By using frontier LLMs to help us label a larger and more diverse set of coding agent sessions, we were able to bootstrap the two models discussed in 1) to a better state, while also providing even richer reward signals for RL. 3) smarter cache-eviction impact calculation. One of the hardest parts of routing well (if you care about saving money) is using the model caches intelligently. We built a subsystem that can calculate the expected value of switching models (and thus paying a high one-time cost to fill up a different cache) much more accurately, helping us avoid costly and unnecessary switches in more cases, while still switching when the benefit outweighs the cost. This is where most of our improvement on cost has come from. We still have a lot of room to continue to improve (we won’t rest until we’re consistently beating Astra/Fable, not just tying!) but matching frontier model performance was a huge milestone for our routing model, and in my opinion validates our initial hypothesis that an ensemble of models can do better than any single model ever could. Our router is open source ( https://github.com/weave-os/router) so anyone can try it out. Or if you prefer you can use our hosted version ( https://weaveos.com/router).
2026-09-25 mikehasa/golive-skill S Take your agent-built product live: hosting, database, domain, email, payments — on your own accounts. Open-source Agent Skill + zero-dependency Node CLI: detect → plan → approve → apply → verify. No GoLive account, backend or telemetry.
2026-09-10 Show HN: Self-hosted company OS, Claude Code and Codex agents in departments S Hi HN, I am Dimitris. This is a company OS that I built and use to run my business and anyone can install it and self host it for free. Think of it as Claude Code, Cowork and the cloud sessions in one self hosted application. It is a Multi-tenant application by design where many people can collaborate on the company agents with 4 different modes of collaboration, and it runs with your Anthropic or OpenAi subscription or even with local models. Every agent can run on Claude Code or Codex CLI running as a persistent process on your server in a kernel sandbox (bubblewrap) with network isolation always on (pasta), with its own workspace, memory, schedules and tools. The same agents can also be configured to run identical on any remote computer through one outbound WebSocket (no inbound ports, no VPN needed). The agent already have lots of built in features, as an example they can answer and place phone calls through Twilio or your own Asterisk, they can edit videos and excel, word, ppt files and preview them directly inside the chat with collabora and many more. Trying it is one install script and a docker compose, no signup. The license is Fair Source, all the code is public and self hosting is free up to 5 users. 1.6.0 went out today. AI Disclaimer: large parts of OtoDock are written using OtoDock itself, running Claude Code. I would love people to read the code, take a look in the sandbox model, and tell me what would stop you from running this on your own hardware. GitHub: https://github.com/OtoDock/oto-dock Website: https://otodock.io
2026-09-03 Shopify/claude-for-commerce-examples S Shopify implementations of Anthropic's commerce-agents blueprint: a storefront shopping agent over UCP and Sign in with Shop, and a merchant agent over the Admin API
2026-09-01 Show HN: 49 IDE – 2D Canvas for Agents S Track agentic CLIs across multiple providers and repos, from any device and across multiple machines, see git trees, terminals, usage, issues, and UI all on one workspace. I have been trying to parallelize agentic coding to extreme, sometimes managing up to 15 CLIs at the same time across multiple repos and tasks, and I suffered a huge productivity tax due to context fragmentation and constantly forgetting which terminal is doing what. I felt restricted by most grid pane managers, and thats why my friend and I built 49 IDE.
2026-08-28 sapientinc/PRAXIST S Autonomous research system for measurable, computer-executable research.
2026-08-23 duty1g/x64dbg-mcp-server S x64dbg-MCP Server is a native MCP (Model Context Protocol) plugin for x64dbg that exposes the debugger's full functionality over HTTP. Connect any MCP-compatible AI assistant and control x64dbg programmatically: set breakpoints, step through code, read memory, dump registers, and more. Built with Zig — zero dependencies, single-binary output, cros
2026-09-10 Show HN: Give your AI agent on-screen guides that show users where to click S Hey HN. I'm Christian, one of the founders of Frigade (YC W23). I've noticed that a lot of in-app AI agents struggle to actually understand the products they exist in. For instance, let's say a user asks an agent how to do something in a given SaaS product. In an ideal case, maybe that agent replies saying it has a tool to do the task and just automates that work entirely for the user. That's a great outcome. But often that's not the case. Maybe there is no tool call for that exact task, or maybe the user's question is best solved by a specific UI workflow or interface. In these cases, many agents tend to fall back on basic RAG on their help center, or sometimes even searching the internet for an understanding of their own product. This can be a very slow process and most of the time help center articles are outdated as products evolve faster than them today. Even worse, no one likes reading a long list of bullets and mapping that back to a UI. My tool (Assist API) solves this gap with a single tool call defined like this: const frigade_guide_tool = { description: 'Call this tool to answer product questions or guide the user through a task.', parameters: { query: { type: 'string', description: 'What the user is asking or wants to do', }, }, run: ({ query }) => frigade.assist({ query }), } Here's a demo I recorded on how to set it up with the Vercel AI SDK: https://www.youtube.com/watch?v=9WQ0UbLjC6I When called, the tool will do the following: 1) Gather context on what the user is seeing on screen, their permissions, feature flags, and more. Then one of the following: 2a) If solvable: Generate an on screen guide for how to fix a given problem 2b) If conceptual: Return text describing to the parent agent how to solve the problem 2c) Reject (i.e. unable to help) How does the tool know what to do? The tool learns a given application UI by using a browser-based agent. You provide a test account to your software (i.e. staging og preview), and a browser agent logs in and works its way through the entire product. It then builds its own map of how the application works which can the be queried about any product-related question or how to get from A to B in the UI. It also writes its own documentation from this map. The agent re-runs on a schedule or can be triggered through CI/CD. Docs and more details: https://frigade.com/assist-api
2026-09-08 EverettFish/holo-card-studio S Turn the user's description or uploaded reference into a finished, editable Blender card and an interactive Three.js page. Preserve the requested subject, style, typography and destination. This skill contains code and text only; generated artwork belongs in the user's output project.
2026-09-03 Show HN: Aura – a Rust agent that investigates and fixes production incidents S We run a SaaS that handles petabytes of data. Our SRE team experimented with using claude, openclaw, langchain, etc. within our incident response workflows. We struggled with overflowing context, lethal trifecta vectors, hallucinations, and burned a lot of frontier tokens mostly on easy work. Approval fatigue was a challenge, and we drew a hard line at relaxing permissions in production. Long story short, we built and open-sourced AURA, a Rust-based harness specifically designed for the type of operations work which routinely involves large volumes of telemetry data and coordinated investigations across many domains of knowledge / state. We have found even on open-weights models, root cause accuracy has been very good, and remediation actions are guarded with human-in-the-loop. AURA runs from a centralized configuration file where workers are defined and scoped to task domains, (e.g. logging review, metrics analysis, and git/scm related queries). All permissions, tools access, LLM backend(s), worker prompts, and the main coordinator prompt are also defined in code. Permitted tool use is enforced deterministically outside the agent's context, so the agent cannot grant itself capabilities through prompting. AURA also handles the rest of the execution layer, such as human approval for sensitive actions and fault tolerance. A lot of thought went into context window management: large tool outputs and worker responses are persisted to disk and agents are given tools to slice/read as needed. The project is Apache 2.0. You can install rpm/deb/brew natively or pull the source and build your own. It runs either as an assistant on your workstation, or as a daemon that you can connect to via any OpenAI-compatible frontend. The GitHub repo is here: https://github.com/mezmo/aura/ The whole thing is free/Free. No signups or open core trapdoor. Demo: We put together an 8-minute video showing a checkout outage resolved by following evidence from several different systems: https://www.youtube.com/watch?v=TERHoRzT8cE tl;dw: 502s from a checkout service cause Grafana to trigger an incident via PagerDuty. AURA (using DeepSeek-V4-Flash) uses its workers and tools to correctly determine root cause as a memory leak in a downstream service, and attributes the incident to a defect in a recently merged PR. The demo then concludes with a human-gated tool call to GitHub, where AURA documents the exact lines of code that are causing the problem, and recommends a fix. Once the PR is deployed, AURA validates that the errors are clearing and transactions are no longer failing. The whole time, AURA’s leaving an audit trail of OTEL events into Phoenix. Technical details: AURA uses an agent coordinator that drafts, executes, and supervises DAG flows through user-defined worker agents to solve complex investigations in parallel. Worker agents are bound to strongly typed durable artifacts used as evidence packets, and the system discourages wasteful recalculation of remote data. AURA also has a powerful header routing system that doesn’t allow the agent to actually touch your keys. HITL interrupts over webhook are first class citizens and have a schema that is easy to adapt to your own workflow. HA deployment options are coming soon. What’s still rocky: - We’re currently working on the async input system for when you’re running AURA as a service. Right now the API just accepts a request, and streams messages until the main loop completes. It’s easy to integrate into a workflow, but the logic required winds up being heavier than it ought to be. - Because there’s no inbound webhook interrupt mechanism, automating AURA for IR requires middleware to invoke it or it needs to poll an MCP for alert escalations. We’re looking for users, contributors, guidance on where to take this next, or just GitHub stars. Looking forward to reading and responding to feedback here.
2026-10-16 feder-cr/open_scion S Open-source version of Scion: version control built for coding agents.
2026-08-27 UditAkhourii/cdaf S CDAF (Cached Descriptive Asset Files) - open sidecar format for video so AI agents stop re-analyzing the same footage. Spec, CLI, agent skill, reproducible benchmark.
2026-08-24 truespar/sentio S Email inbox API for AI agents. Give every agent its own real email address, receive mail as structured webhooks, and reply in-thread over REST. A complete multi-tenant mail server in Rust: inbound and outbound, DKIM/SPF/DMARC/ARC, MTA-STS, DANE, three-tier anti-spam.
2026-10-12 NousResearch/hermes-official-plugins S ☤ The official plugins for Hermes Agent, built and maintained by Nous Research: Home Assistant, Spotify, Claude Subscription DirectSDK and more, one install away from the Hermes plugin catalog.
2026-10-04 Show HN: Offrun – manage every coding agent from one workspace S Run Claude Code, Codex, AGY, and Grok Build side by side. See who is working, who needs you, and what every account has left.
2026-09-30 Show HN: Raven – The harness of harnesses, built for RSI S
2026-09-27 Show HN: Reladraw – A diagram language where you decide where to place things S I love making diagrams to help understand, plan, etc. However, the options are (A) auto-placement languages like Mermaid or Graphviz (which don't let me decide how the diagram looks), or (B) software like Draw.io which are powerful but are very time consuming (and inefficient for agents to manipulate). I wanted to have the benefits of both, where you can define a diagram in a diagram language, but also retain a high degree of control over what the diagram looks like. I also wanted this to work well for humans and agents. On the Github link, there's a playground where you can try it out without installation. There's also instructions for a simple npm install and for installing a skill you can use with Claude or other agents.
2026-09-03 dreamers-laboratory/agent-fleet-manager S A general-purpose engine for large-scale, repeated information gathering by a fleet of workers.
2026-08-31 useagenthq/useagent S Hand off the work. Get back the result. The open-source AI coworker for your team: agents with their own cloud computer, your tools and context, handing back finished work - websites, decks, spreadsheets, reports, PRs. Runs Claude Code, Codex, OpenCode on your subscription.
2026-10-19 alchaincyf/huashu-motion A 让你的agent一键直出商业级产品宣传片：会作曲，能指着画面改。Commercial-grade launch films by your coding agent — point at the frame to change it.
2026-10-15 outcrawlai/scout A Scout: the open-source browser agent, by Outcrawl. Mirrored automatically from Outcrawl's own repository.
2026-09-11 Show HN: Open-source simulation testing infra for voice agents A Hey HN, we’re Nischal & Naman. We’re brothers, and together we’re building an open-source platform for simulation based testing of voice agents (try it out in 5 mins - https://docs.egma.ai/docs/get-started/quickstart, 2 min demo video - https://youtu.be/wgDWEe5UAUY) Platforms that help you do simulation testing already exist. But they all charge a heavy premium on top of inference costs. We believe if the industry truly wants to scale simulation testing of voice agents, we need to stop charging a premium on inference and start providing infrastructure to scale simulations. The project is built in a way that allows you to bring your own STT, LLM, and TTS provider keys while also providing a way for us to provide inference directly within the platform at a 0% markup so you don’t have to wrestle with multiple keys/ rate limits. A bit on why’re we’re building this - We started working on voice agents in early 2024 and since then, have worked on numerous voice ai systems - like screenless voice-powered hardware for kids[1], AI receptionist deployed in healthcare practices like med spas & therapy clinics and personal accountability coaches. Some of these were full-fledged startups others were just side projects. But we kept on encountering similar issues all throughout. One of the most frustrating parts was calling the agent again and again, reciting the same script to test its behavior. We also kept on encountering new issues all the time in production that we couldn't have simulated pre launch. This frustration led us to a bigger question: how can developers trust the voice agents they’re shipping? We believe building that trust takes two things - - First, you need a way to test the major scenarios your agent will face in production before you release it, without having to make every call yourself. But you can’t test for everything. Real world is too messy to predict in advance. - Which brings us to the second: you need a way to find issues once your agent is in production, whether it’s handling a few dozen conversations or millions. Detect drift in known behaviors AND surface unknown-unknown ways in which your agent is going wrong. We think it’s a really hard problem to solve. And having felt it firsthand, we’re deeply motivated to take a shot at it. Today we’re launching the simulation testing side of the platform. We feel it’s mature enough that real teams can depend on it. For eval design, we took inspiration from anthropic’s evals design[2] and extended it to voice systems. For technical & business model design we borrowed ideas from Langfuse[3]. Our stack is postgres, clickhouse & minio. Its easy to self-host[4] & the code has a permissive MIT license. We also have managed cloud version. We’ve written more about our [testing]( https://docs.egma.ai/docs/core-philosophies/testing-philosop...) and [monitoring]( https://docs.egma.ai/docs/core-philosophies/monitoring-philo...) philosophy in the docs. We’d love feedback from the HN community and people building voice agents - how are you testing today, what’s working and what’s frustrating? We’ll be in the comments. Thanks! [1] https://x.com/theBhulawat/status/1966200231705595932?s=20 [2] https://www.anthropic.com/engineering/demystifying-evals-for... [3] https://github.com/langfuse/langfuse [4] https://docs.egma.ai/self-hosting/get-started
2026-08-29 XiaoDuoYa/codex-with-chatgpt A ChatGPT thinks. Codex works. Use ChatGPT as the planning brain while keeping the Codex harness.
2026-08-24 liangdabiao/deepseek-v4-flash-vision-video-rag A DeepSeek V4-Flash Vision Video RAG 让 AI 真正"看懂" 一段视频，然后你对它提问：它告诉你答案、答案发生在 第几分几秒，并切出那一段的可播放片段和关键帧给你核对。 基于 DeepSeek 视觉大模型 deepseek-v4-flash-vision-exp 的视频理解与问答 （video RAG）agent skill。先按时间轴抽帧阅读、建立索引（一次性），再对问题做 本地粗筛 → 视觉精排 → 深读回答；回答带 [MM:SS] 时间戳引用，自动生成 自包含 HTML 预览页（内嵌可播放片段 + 关键帧 + 答案），双击浏览器即看。
2026-09-29 Show HN: OpenAPPA – open-source deterministic guardrails that don't break agents A Hi Hacker News! Matvey, one of the authors, is here. While building enterprise agents, we ran into a problem: the more tools you connect to the AI, the higher the chance it will run out of control and leak sensitive data. Guardrails, in theory, should prevent this, but the situation is worrying: - Non-deterministic guardrails (LLM as a judge, auto modes, etc.) are vulnerable to prompt injections, or they lack knowledge of the data, making them inefficient (~10% data leaks on our benchmarks). - Existing deterministic guardrails (Cedar, OPA, FIDES, Dogwood) require massive case-specific IF-ELSE-like policies and break agents (~59% utility loss on our benchmarks). We did something differently. We’ve taken the best of existing deterministic guardrails and built a policy language that is data-specific, not use-case specific. It lets you scale agents without updating a policy. On top of that, we’ve added multiple tricks (like a remedy plan or a DualLLM pattern) to help agents operate within those restrictions, raising utility from ~40% to ~90% and making it the first deterministic guardrail that doesn't break agents. Finally, we’ve designed it to be pluggable into any agent loop with pre- and post-tool-call hooks. We invite you to check out our benchmarks: https://www.openappa.com/evaluation Play with it in Claude Code: https://www.openappa.com/claude-code Try plugging it into your agent: https://www.openappa.com/add-to-agent Or check the academic paper: https://arxiv.org/abs/2607.24625 We'd love to hear any feedback!
2026-09-25 Show HN: PlaceCall (YC W26) – agentic API to call businesses and get things done A Hi HN, this is Yarik and Vlad from VOYGR - we are building the tools for agents and apps to engage with local businesses. It all started with our own pain point at VOYGR: calling businesses to verify if they are open. We are both from Google (Maps and Search) and even there, the merchants and venues don’t keep this info updated. So we built an API and started using it in-house. On July 4th, we were driving through Portland looking for a place to eat. Google Maps was saying “Holiday hours may vary”. We tried some AI assistant apps (Poke was trending back then), and they couldn’t help. Our API found 15 Thai and sushi places and called it in parallel. A few confirmed that they are open today and are taking walk-ins. One was marked closed on Google. The app can find the restaurant, but someone has to answer the phone. This is the problem we kept seeing. Assistants can’t do much in the real world where the phone number is still a major communication interface. Muse added calls in beta mode, but it’s still calling only one place at a time. There are a bunch of tools to create your own voice agents, but nothing that can work simply out of the box. We’ve built PlaceCall to fit all your agentic needs. Give it a task in plain English and a US phone number (or a list of numbers) - it navigates phone menus, waits for you on hold, and returns a structured response with outcome, recording and transcript. If you don’t have a phone number, just describe what you need and PlaceCall comes up with the recommendations. Here is the demo: https://youtu.be/7ZbVCPv8Wq0?si=LoLxxx1M7nhS7_j- When PlaceCall API was tested in our hackathon, we had the following use cases implemented by the developers: checking if a doctor accepts new patients, getting quotes for car service, and running quick surveys. The API can be used through Claude Code and Codex; examples and setup are listed in the repo: https://github.com/voygr-tech/placecall We give the first 250 calls free. Product page is here: https://voygr.tech/placecall/ Try it and let us know how it goes. We are interested in learning what you would use it for and what is missing. We are here all day and happy to answer any questions. Our email: [email protected]
2026-08-29 JordyZomer/lemmalog A A Datalog engine for LLM agent memory: stratified rules, provenance-tracked facts, incremental derivation, and an MCP server that lets your harness use it as a shared brain.
2026-10-19 Show HN: Sphinx – A hook that quizzes you on your AI's code before shipping A FOSS, Basically it plugs into your existing coding agent workflow and every time you push to prod it'll throw up a questionnaire that checks if you understand how the code you're pushing works.
2026-10-17 anujasid/blender-ai-mcp-addon A AI assistant and MCP server for Blender: visual chat, scene tools, multi-agent workflows. Works with OpenAI, ChatGPT, Claude, Gemini, Ollama, OpenCode. Blender AI addon and plugin, copilot for 3D, bpy Python scripting, MCP integration.
2026-10-10 robbin/android-adb-control S An Agent Skill for Android phone control via ADB: observe, act, and verify. For Pi, Codex, Claude Code, and OpenCode.
2026-10-10 Show HN: Let your AI agents paint big arrows, boxes and text on your screen A
2026-09-27 feitangyuan/onetake A Motion films that never cut to the next slide: every beat grows out of the one before, one continuous camera, continuity measured by an oracle. A Claude Agent Skill for product launch films and feature demos.
2026-09-27 graygnatconsole/mcp-audit-tool A 🛡️ Security audit CLI for Model Context Protocol (MCP) servers — scan AI agent configs for tool poisoning, rug pulls, hardcoded secrets, command injection & supply-chain risks. Pure Python, SARIF + CI ready.
2026-09-21 zai-org/ZCode A Z.ai's coding agent harness. Powerful, intelligent, extensible.
2026-09-15 Matthew0822/ToolReplay A Audit AI agent tool-call transcripts: hash-chain sealing, deterministic replay, and scope overreach checks. Dependency-free Python CLI.
2026-08-29 usedotai/dot-reflex A Open-source agent execution-recovery controller for coding and tool-using AI agents.
2026-10-10 franzenzenhofer/big-arrow-on-the-screen A Let your AI agents paint big arrows, boxes and text on your Mac screen. One CLI, click-through, gone by itself. Skill for Claude Code and Codex. MIT.
2026-10-06 Show HN: Minigraf – An embedded, bi-temporal graph database in Rust A A successor project to my previous attempt at building a temporal graph database ( https://news.ycombinator.com/item?id=23455516). This one builds on its predecessor's concept, but drops the dependency on another database to provide the storage, transaction and query engines. Minigraf is self-contained, packaged as a library crate, embeds into your application, and supports full bi-temporality.
2026-10-01 rehan-remade/universal-modder A Point Claude at any game. Skills, tools and the fal MCP that let Claude Code mod almost any PC game you own: recon, reverse engineering, fal-generated art/3D/audio, in-game testing, showcase videos.
2026-09-12 Hona/openeval A Typed prompt-and-judge SDK, CLI, and evidence viewer for agent evaluations
2026-09-10 Show HN: Type.com: Multiplayer Codex/Claude in the cloud for non-tech use cases A Hey HN, I'm Komran, the CTO and cofounder of type.com. We're launching today. I've been describing type as "the IDE for non technical people." It allows you to build really cool shit without having to worry about maintaining the perfect local dev environment. Everything is cloud first by default. No fumbling with handoff flows or anything weird like that. Every thread is backed by a persistent sandbox VM so you don't even have to worry about git worktrees if you're doing something git backed. You can start work from claude, codex, slack, or email and then push it into a shared session in Type where your whole team can co-prompt. Since our "meta-harness" is a wrapper of codex-cli and claude code, you can connect your existing subscriptions. Because we're a multiplayer product we had to make sure that we only bill the subscription for messages authored by that user. If Alice has a subscription and she starts a thread with an agent, she will have her subscription used. If Bob comes in and add their own prompt but Bob does not have a subscription, they will be billed against their type usage. We don't charge any margin on tokens. Even with how early the product is, our customers love us and prefer us to using claude code. If I had to say our advantages over claude tag/claude cowork it would be these: * You can switch models/harnesses anytime, you aren't locked into a single provider. Creating a custom internal app/artifact is better on type because we handle internal hosting, RBAC, version control, dev/prod environments, and even the ability for your apps to talk to your type integrations. Sharing an internal app is very akin to sharing a google doc. * It's just claude code and codex under the hood. In general it should work very similar to how codex and claude code work * At my last company, halp.com (acquired by Atlassian) we invented the term "Slack-first" because we knew that people loved doing work without context switching. Type is great in slack. Each slack channel maps to a public or private "Space" which has access read only or read/write access grants to integrations. Each space has a default harness, model, and effort level built in. So you can use your subscriptions all from slack. * Our CLI allows you to get the best parts of type even when you're working in single player mode on your local machine. The type-cli lets you push your local session up to type so that other can take a look at it. It can also push up a doc for collaborative editing, and even reuse the same integrations that you've been given access to on type locally. You can sign up today at https://type.com. You get $100 worth of credits to mess around with it. It's been really fun building this. Competing with the likes of grok bot, codex, and claude cowork has been really fun, but I think that they are still very much in the personal assistant, single player mode. Each person has to set up the same integrations, skills and automations need to be created all from scratch, and every frontier's product is gonna make you be locked in to their model.
2026-09-09 daishuge/pcb-skill A Agent skill for taking a hardware idea to a manufacturable PCB: concept, schematic, sourcing, layout, routing, verification, and a purchase staged to the pre-payment page. Drives EasyEDA Pro over MCP from Claude Code or Codex.
2026-09-08 achimala/dream-loop A Agent skill for impressive 3D visuals using Blender + image gen + subagent critic
2026-10-15 RuihongY/verdi_bridge S Connect AI agents to a live Synopsys Verdi session through MCP: waveforms, signal values, driver tracing, and RTL source.
2026-10-08 ostiums/fixkit S Long press any element of your iOS app in the simulator, say what is wrong, and the report lands in your Claude Code session
2026-09-23 malevrigns/agent-jev A AgentJev-0.6B - a fast 'System One' decision model for AI Agents: feed it any unstructured state (diffs, traces, logs) and structured questions, get calibrated probability distributions back in one ~50ms forward pass. Zero output-token decoding.
2026-09-03 gozen3ji/consulting-pptx-skill A AIにまじなPPTXを作らせるClaude Codeスキル — スライド規約＋38型SlideSpec＋生成パイプライン＋機械チェック
2026-10-07 longsurf-ai/openchart A OSS Tradingview with your own AI, free and unlimited.
2026-10-07 Show HN: OpenChart – OSS TradingView alternative with your own AI agent A Hi HN, I'm Weilun, cofounder of OpenChart. We built OpenChart because we wanted Claude and Codex to interface with the markets, like most of humans do, through charts rather than through CLIs. With your existing AI plans, your agent can work directly with your charts: annotate a setup, write an indicator, investigate a market move, or create an alert. The chart, conversation, and research live in the same workspace. OpenChart alerts on any market move: draw any shape on a chart and attach an alert. When price crosses it, the alert triggers your agent to investigate the move and save the research in your workspace. The agents can work with charts, indicators, and alerts. Custom indicators and alert conditions are written in [Tea]( https://github.com/longsurf-ai/tea), a programming language we built for the markets. We believe software shouldn’t impose artificial friction on your workflow. OpenChart is built on that belief with no software cap. The source code is publicly available, and you can use free market data from day one. We also offer optional paid cloud market data for lower latency. Everything is private. Your charts, research, and indicator code are stored on your own computer, with the app's backend running locally too. We currently support only macOS on Apple Silicon, but we're expanding to more platforms. Website: https://openchart.co/ Download: https://downloads.longsurf.ai/openchart/darwin/arm64/OpenCha... We built OpenChart because we want to explore how assets can be co-manages by agents and humans. I'd especially like to hear how you are using agents to help with your investment journey. If you like what we're building, a star on GitHub would mean a lot to us: https://github.com/longsurf-ai/openchart
2026-09-30 mlc-ai/TIRx-harness A An Open Compiler Harness for Agentic GPU Programming
2026-09-25 Show HN: Canary (YC) – Independent verification for AI code A Hey HN, we are Aakash and Viswesh and we are building Canary ( https://www.runcanary.ai/) - independent verification for AI code. Claude/Codex calls Canary with the changesets, intended behaviour and team knowledge. Canary then deploys agent swarms to investigate potential failures and test suspected runtime bugs in remote sandboxes. To try it on your repository, paste this into your coding agent: Install the Canary CLI with npm i -g @runcanary/cli, then run canary skills and follow its instructions to onboard this repository. Verification starts with what software is supposed to do and most importantly what it must never allow. This means investigating how inputs, permissions, state, timing, dependencies etc interact with each other. Intent is not always fully declared as well but many expectations are clear: private files should stay private, credentials should not leak, and retries should not create unintended duplicate effects. We believe the future is a unified and independent verification system that starts with all those expectations and then chooses how to investigate each suspected failure. Source-only code reviews catches static issues in the implementation but even a clean review leaves a good chunk of behavioral only issues untested. Unit tests, integrations, E2E, static analysis, runtime experiments and formal verification are all means to establish that behavior thereby generating different kinds of evidence and guarantees. This is why we believe a dedicated verification harness that can think and reason through all these modalities and invariants is necessary on top of general intelligence. The harness needs to start with the system’s intended behavior, develop a series of potential failure scenarios and choose how to investigate them. It’s sole functionality is to pressure test and challenge the assumptions behind a change, create the conditions needed to test suspected failures and assess what the resulting evidence establishes How Canary works: it takes a cold snapshot of the codebase when called, combining the supplied intent and team knowledge with requirements, decisions, prior issues from tools like Notion, Linear. It can also route questions to you through the coding agents if anything is ambiguous. Canary’s harness coordinates agent swarms by leveraging the different strengths across model families. It compares the code before and after, traces the effects through callers, dependencies, state transitions etc. and each suspected failure becomes a concrete scenario with an actor, state, trigger, outcomes and many more runtime states., For each suspected failure, Canary chooses the best way to provide evidence through methods like runtime verification, static analysis, unit, integration or sometimes even combination of these as necessary. The agent executes these checks in remote sandboxes by seeding data, configuring permissions, mocking dependencies and third party integrations and much more. Canary then returns these findings and supporting evidence back to the coding agents which then fixes these failures and requests reverifications against the failed scenarios. To get started, give your coding agent this setup instruction and tell us what it caught and how we can do better. Install the Canary CLI with npm i -g @runcanary/cli, then run canary skills and follow its instructions to onboard this repository. We are still pretty early in our journey and would love feedback on the product and how we can do better.
2026-09-16 Show HN: Pizza Bot – An inbox for AI agents that work in the background A Hi HN - long-time lurker (since 2012!), first time poster. Pizza Bot is a self-hosted desktop app for Mac, Windows, and Linux that runs AI agents in the background and exposes them through an email-like UI. Finished work shows up in Unread, and anything waiting on your approval shows up in Action. It's Apache 2.0-licensed, there's no signup and no telemetry, and you bring your own model provider: Anthropic, Amazon Bedrock, Google Gemini, OpenAI, OpenRouter, or a local model through Ollama. There are builds on the releases page, or you can run it from source. Pizza Bot started as an internal passion project I worked on with a small team at Amazon. The whole thing came out of my frustration at having to manually log CRM activities through a browser form. I built a simple REST API called "JoeBot" that connected to my authenticated browser session over CDP and filled out the form for me using Playwright. Then I hacked up a quick Obsidian plugin so I could trigger it from my local notes (no AI and no MCP servers involved). This caught on quickly. My fellow AWS Solutions Architect Igor Fil joined up with me, and we rebranded the project as "Pizza Bot," named after Amazon's two-pizza teams. We started seeing what other automations we could build. We found a GraphQL API we could query and hacked up some "recipes" to pull data out of the CRM to help with meeting prep. That worked great, and it was right around the time MCP servers seemed to be taking off, so we decided to expose Pizza Bot as an MCP server instead, so it would be available to AI tools through natural language. This was a decent solution for technical users, but the Account Managers who live inside our CRM system wanted something too. We decided to rebuild Pizza Bot as an Electron desktop app modeled after an email inbox, so it would be familiar to non-technical users and would run on both Mac and Windows. We also bundled internal MCP servers as OCI images and hosted them in Amazon ECR as an "addon marketplace" so users could install them with one click without having to set up Amazon developer tooling. The project took off organically and expanded outside of AWS into the wider Amazon organization globally. More than 2,000 people ended up using it for meeting prep, email drafting, Slack summaries, CRM logging, prioritizing their day, and web research. Once apps like Claude Cowork and Amazon's own Quick Desktop came out, we realized the real growth opportunity was outside of Amazon. Rather than try to rip out the Amazon-specific integrations, we rebuilt Pizza Bot once more as an open source project. We leaned on coding agents heavily, which is the only reason a team our size could pull off a full rewrite. I'm pleased to say it's finally public, and we're hoping to bring in community members and see where it goes. We'd like to do for knowledge workers what Claude Code and Codex have done for programmers. A couple of things to know up front. Most of what made Pizza Bot useful on day one inside Amazon came from that internal catalog of skills and MCP servers for Amazon's own systems, and none of it could come out with the app. So it ships thinner than the version those 2,000 people used, and building that catalog back up for tools other people actually use is where we need the most help. It's also a community project and not an AWS service, so there's no support or SLA behind it. The Windows and Linux builds aren't signed yet either. On the technical side, Pizza Bot is a server and a client. The desktop app bundles both, or you can point a client at a remote backend; personally, I self-host the server on my home network and reach it from my phone over Tailscale. The server owns the thread lifecycle and checkpoints state with DeepAgents and LangGraph, and clients rehydrate from it as needed, so you can disconnect mid-run and pick the thread back up from another client. Approval pauses outlive the session that created them and collect in an Action filter, so you can answer an hour later from a different device. The agent you talk to has a sandboxed QuickJS interpreter that can reach your filesystem only if you grant it a folder, but its main job is to delegate. Each subagent is a 1:1 mapping of a Skill, and an Activity bar shows that subagent and the tool calls it's making as it works. Memory is opt-in and stored as plain markdown files on your machine. Every tool call is explicit, including looking up a memory - we err on the side of transparency to reduce surprises. Tools come from MCP servers, and skills are ordinary SKILL.md files with a per-tool approval policy, so existing skills that don't require a code interpreter should still work. What I'd most like to hear about is where the app itself gets in your way, the kind of problem you can't fix by writing a skill or an MCP server. I'm around today to answer questions!
2026-09-13 Show HN: Graphify C# – Compiler-accurate Find Usages for coding agents A
2026-09-06 Show HN: Moadim.io – A scheduler for agents A Why can't we get an agent scheduler that supports all of the following: - git compatible - agent agnostic - 100% open source - os and system-agnostic - multi-runner support - support mcp/ui/http - unlimited routines/crons So I built one, moadim.io is a local Rust daemon you install in the target machine, give it a name and manage its routines via a Git repository, wants a new routine that send you a daily message from this machine? Create a pr and merge, have another routine that run every hour to pull the latest changes to the ~/.config/moadim folder. With more than 1,000 users, I define this project as almost "done" and ready for production. Me and thousand more people are use it in a daily manner. It currently supports Claude, Codex, Hermes, and Pi, and you are welcome to add your agent of choice as well because it's 100% configurable. You are welcome to have a look at the source code of the daemon in " http://github.com/moadim-io/daemon" Feel free to provide me with suggestions for more features around this topic. I don't want to branch out to new off road topics likt webhooks, this is a "done" software in the realm of agents schedulers that focus on cron-like work. Also feel free to start the github repository and open issues and PRs for your suggestions.
2026-11-08 lxe/phoneuse S See and operate an unmodified iPhone from a Linux PC: AirPlay mirror in, Bluetooth HID mouse/keyboard out, plus a CLI and skill for agents.
2026-10-21 Umair/stalemap S Tell your coding agent which dependencies changed after its training cutoff. Diffs types and changelogs, writes a short block into AGENTS.md / CLAUDE.md.
2026-10-20 vercel-labs/fx0 A A lightweight design agent built with libfx. Describe a page and an agent writes it as one HTML file.
2026-10-09 mhtsec/ARTEX A AI 自主渗透测试系统 | 百度“agent+”攻防挑战赛冠军项目
2026-09-12 FankChen/tracecrate A Local-first AI agent trace workbench. Inspect Claude Code, Codex and OTLP logs, compare runs, and export privacy-conscious reports. No backend or API keys.
2026-10-17 reason-machines/reason-network A Reason Network: your team's coding agents, talking to each other. The engine behind `reason network`.
2026-10-14 Show HN: SolidCraft – Open-Source Parametric CAD in Rust (With MCP Support) A
2026-10-14 Show HN: Agor, multiplayer AI – stop running agents alone A
2026-10-13 Show HN: Atticus – Hold-to-talk voice assistant for macOS that runs on Claude -p A Hi HN! I'm Saarth. I'm in 9th grade, and I built Atticus. Atticus is a fast and cheap Mac assistant that can do tasks for you. You hold down Control-Option-Space, speak or type your task, be it complex or a simple browser search, and Atticus will get on it. The interesting part is that simple tasks, like opening apps, playing a Spotify playlist, or web queries, don't touch AI, and instead are processed locally and typically finish in under a second. For example, prompting Atticus to open apps takes around 0.08 seconds, sometimes even finishing before you finish your sentence, sort of like Jev. Any tasks that are more complicated will be sent to Claude via claude -p, meaning you never need an API key and Atticus will run on your existing Claude Pro or Max plan. By using this approach, costs are also cut down immensely, with, on average, tasks costing a fraction of a cent. To achieve Atticus' speed, it remembers shortcuts from each task it completes. When it does a task right for the first time, it saves its steps and replays them the next time without AI. For example, a workflow of creating a new meeting took about 30 seconds the first time around. Atticus then logs the exact steps to memory and takes 5-8 seconds on the next pass. To add to this, Atticus also learns about you as you use it more and from what's on your computer, including your name, your apps, and things you want it to remember, personalizing its answers to you. To combat the safety concerns over giving an AI unrestricted access to your device, Atticus has toggleable restrictions, so before anything is deleted, sent, bought, or written to a file, the user is prompted before that action. Some features I'm currently working on include a meeting copilot that capitalizes on Atticus' speed, allowing Atticus to listen in on meetings and answer questions in real time as they come in, personalized to the context of the meeting and what it knows about you. To run Atticus, macOS 26 or higher is required for its voice transcription features. The voice transcription is also done locally by macOS. Atticus is free, open source, and also very fresh, and I'm completely aware of its many rough edges, which is why I'd love to hear from everyone about what breaks and what you guys would like to see Atticus do!
2026-10-01 spenmcke/compress A Free CLI from Everest that compresses Codex tool output.
2026-09-21 shhivv/arc-cua A Superfast action layer for computer-use agents
2026-09-18 Show HN: MCPJam - the first testing & evaluations platform for MCP servers A Prathmesh, CEO of MCPJam here. Users now start in ChatGPT, Claude, Cursor, and other AI clients. They reach your product through your MCP server. That means your users often aren’t in your product anymore. You can’t see what they prompted for, how the agent interpreted it, or whether your server helped them get the result they wanted. I saw this firsthand leading MCP technical strategy at Asana, including our ChatGPT and Claude launches. We were building high-stakes enterprise integrations, but we had no reliable way to test them the way we test normal software- or to know whether they worked once they reached real users. I started using MCPJam for those problems after re-connecting with my former coworker who created the project. brought it to more of our developers, and worked it into our CI/CD pipeline. I joined the team because I kept hearing the same issue from other companies building for agents. So, what does “good” look like for MCP? For us, it means users reliably get the outcome they came for, across the AI clients they use. That’s what we’ve been building toward. MCPJam now helps you test the full workflow, from the first prompt to the expected result: * Swarms: Simulate users with different goals and prompts to find where workflows break across AI clients. * User Testing: Watch how real users interact with your MCP product, where they get stuck, and how they feel about the results. * Evals: Turn those workflows into repeatable tests that check whether users get the expected outcome. * CI/CD: Run those evals across AI clients before each release to catch regressions. Over 106,000 developers and +300 enterprises use our open-source solution to see how their servers behave locally across major AI clients. MCPJam has grown from a debugging tool into a continuous testing and evaluation workflow for MCP servers. If you’re building an MCP server or agent-facing product, give MCPJam a try. What is the hardest thing for you to test? We love hearing about your MCP server builds!
2026-09-08 Albertchamberlain/Awesome-OKF A OKF (Open Knowledge Format) — curated catalog of tools, plugins, skills, proposals, and docs for agent-friendly knowledge. YAML-driven, agent-searchable, MCP-ready.
2026-09-06 Rion-Wu-tech/wechat-intelligence-hub A Local-first WeChat intelligence system with a read-only CLI, Codex skills, searchable chat history, daily briefings, follow-ups and opportunity tracking.
2026-08-28 brooxiacaigonv/AIrecall A Drop-in long-term memory layer for AI agents - episodic + semantic memory, hybrid retrieval, and auto-summarization. Python SDK + Go memory server.
2026-10-11 rociiu/talorys A Your personal AI agent in your own Cloudflare account. Chat, memory, tasks, notes and scheduled reminders, deployed with one command: npx create-talorys@latest. Free-tier friendly, single-user, no telemetry.
2026-09-17 Show HN: I built a router for agent tools A
2026-09-05 sudoingX/nopasswd-sudo A Time-boxed passwordless sudo for agents. Grants expire on a systemd timer, validated with visudo before install, audited with sudoreplay.
2026-09-02 boringmarketer/meta-ads-skill A Teach your AI coding agent to run Meta ads: API setup, programmatic campaigns, browser control, and conversion tracking that's actually proven to work.
2026-08-30 cbrock84/headcount A An agent organization for Claude Code, structured as a company — 15+ departments, 125+ skills, each independently installable.
2026-10-21 browser-use/rl-env-use A Build RL environments for any domain from real cases that Browser Use agents mine from the web
2026-10-17 Show HN: Gemini 4 Argon plays StarCraft A I've refreshed StarSkirmish Bench, a benchmark where LLMs play StarCraft BW by writing code, with results from the flagship models of OpenAI, Anthropic, Google, and Muse
2026-10-17 Show HN: Things to Do API A I built an api/mcp for local events: pop-ups, one-off events, concerts, etc. currently covering the SF bay area. I've been building a bunch of consumer apps lately around personalization and recommendation and have had a hard time finding this data in one spot. agents can find useful stuff through tool calls / web search, but slow. you can scrape on demand too, but i wanted a faster data set and couldn't find an existing provider for the mix of event types I was searching for. the methodology for it is to find influential accounts/posters/etc. and do a fan out across things that they co-post with. mcp: `claude mcp add --transport http thingstodo https://api.thingstodoapi.com/mcp` I'm considering adding other cities. If you're building something that needs local event data, would something like this be useful and where would you want coverage?
2026-10-04 QingYunA/answer-me-with-html A Answer me with HTML — an agent skill that answers hard questions with a one-page HTML you can actually read. 让 AI Agent 用一页 HTML 回答复杂问题。
2026-10-02 Remocn/remocn-studio A Agent-native video editor
2026-10-02 Tuskira/ai-agent-gateway A AI Agent Gateway - an open-source gateway that secures, governs and observes AI agents' MCP tool calls and LLM traffic
2026-09-24 yetone/magpie A Every agent's model. One place. Codex on DeepSeek, Claude Code on Kimi, from the menu bar.
2026-09-18 Show HN: Scry, programmable internet search w/ congestion pricing A Meet Scry, a 500 TB NVMe internet index in ClickHouse that you can run ~arbitrary readonly SQL and some of Datalog over, and I handle the problem of resource-contention with congestion-based micro-auction pricing. When there's capacity, the service is free for non-commercial use. --- Hello. It's 2026, we're training simulated fruit fly brains to play Beat Saber, do we still have to be stuck with internet (re)search as fn: natural language -> black box we can't do anything about -> ranked_list/summary? There is a long history of people trying to do very fancy things that end up being done in relational databases and a little SQL. There is a gravity to them, a bitter lesson, just like scaling of generalized ml training methods. I mean many, many information products can be built off essentially giant real-time OLAP databases and frontier LLMs writing brilliant SQL+Datalog+vector+Jev etc. queries. Google Search, Tavily, Exa essentially have the problem of mapping your agents' context you are willing to provide, to a tiny subset of their index. You pay a fixed cost to an extremely hard problem that has a distribution of hardness, which means YOU eat the downsides when they are running out of budgeted compute to help you out. Their algorithms are opaque to the caller, there's really not much user control, and there's not a serious opportunity to communally improve search recipes, like the lexical+Jev recipes you trust to select bleeding edge AI builders. Furthermore, search companies aren't even pursuing text-to-SQL anymore (several have talked to me)... they made up their minds during the traumatic 2024 text-to-sql days. They were just too early. I hope you enjoy. I'm intent on scaling this paradigm on differentiated hardware over much more data, so any compelling use cases or queries I could show off, would be much appreciated!
2026-09-09 Omni-Interaction-Gander/Omni-Interaction-Agent A Open-source end-to-end omni interaction agent for natural full-duplex voice collaboration, continuous audio-visual perception, and asynchronous long-horizon agentic execution.
2026-09-02 Show HN: Orthogonal – One integration for agents to discover and pay for APIs A Hey HN, we're Bera and Christian, and we're building Orthogonal. Orthogonal gives AI agents one integration to discover, access, and pay for APIs. Think OpenRouter for the APIs that AI agents use. We currently have 50+ API providers and 700+ API endpoints across company, people, financial, and web data, available through our MCP server and SDK on a pay-as-you-go basis. We started with GTM related APIs and are now expanding the catalog. You can try at https://www.orthogonal.com/. You'll need to create an account to use the chat or connect the MCP/SDK. An example prompt for the chat or MCP: Find AI infrastructure startups that recently raised a Seed or Series A and return their founders, LinkedIn profiles, and relevant company data. When we heard about https://x402.org/, we saw an opportunity to make APIs easier for agents to pay for. Instead of an agent needing to manage an API key and billing for every API, what if it could just pay for each request? So we started onboarding existing API providers onto x402. Then we realized an agent cannot pay for an API it doesn’t know exists. It also needs to figure out which API can complete the task and how to use it. We also found that putting every small API payment on-chain did not always make sense. If both sides already have accounts with Orthogonal, settling internally is simpler. So we separated API discovery and access from the payment rail. Today, payments can happen through Orthogonal credits, x402, https://mpp.dev/, or https://www.circle.com/nanopayments. The problem we're working on now is API selection. If five APIs can answer the same request, which should the agent use? The cheapest might have worse coverage. Another might be more accurate but slower. We're working on using price, latency, reliability, and result quality to help agents make those decisions. How are you currently giving agents access to paid APIs? And if you gave an agent a budget, would you trust it to choose which API provider to spend it on?
2026-08-31 Show HN: Cogram Studio – CAD and BIM workspace for humans and agents A Hi HN, Rick and Alex here, co-founders of Cogram. We’ve been making project-management software for architects and engineers since 2023, and are now experimenting with a second product. Cogram Studio (studio.cogram.com) is a CAD and BIM workspace for AI agents to create three-dimensional models and dimensioned drawings. Studio runs FreeCAD ( https://github.com/FreeCAD/FreeCAD) 1.1 headlessly, using the OpenCASCADE ( https://github.com/Open-Cascade-SAS/OCCT) geometry kernel. It imports and exports STEP, IFC, STL, DXF and FCStd files. Studio exposes an MCP server, so you can operate it from Claude Code, Codex, ChatGPT, or other agents. Tools cover scripting against FreeCAD, inspecting the model visually, CRUD operations on views, tables, and sheets, importing terrain, etc. Studio also provides a built-in agent based on Pi ( https://github.com/earendil-works/pi) and a basic UI for 3D model inspection, measurement, view, table, and sheet creation. There are two ways to try it: 1. Bring your own agent. Point Codex or Claude Code at studio.cogram.com/skill.md and ask it to follow instructions there. This starts a one-hour Studio session and opens the modelling environment in your browser, connected to your agent. You need to confirm sign up only if you want to save the session. 2. Use Studio’s built-in agent. Submit a prompt at studio.cogram.com and sign up. You receive 50 free credits. Here are two models created by agents: - A simple bookshelf ( https://studio.cogram.com/view/shr_1agb21nd4) with two sheets - The Colosseum ( https://studio.cogram.com/view/shr_3tm32c64y) [large, can take minutes to load], with two sheets showing dimensioned and annotated views. For now, Studio works best when you use it iteratively. If you ask an agent for a finished, complex model and “let it rip”, it may produce something that looks convincing but doesn't hold up to scrutiny. Agentic modelling feels roughly like where agentic programming was in 2024: useful for speeding up repetitive or bounded tasks but not yet reliable enough to complete long, complex jobs. Recent models have improved at CAD. We expect progress to accelerate as models become even better at reasoning across the different representations used in architecture and engineering: geometry, drawings, images, materials, physical constraints, specifications, regulations, code, etc. Early users and our own experience suggest that agents are already useful design collaborators. They can help with repetitive CAD work, fabrication options, part sourcing and cost estimates. We expect that as models improve, agents will be able to explore many different designs, run simulations, and check constraints across an entire project. CAD will become a shared workspace for humans and agents. Because design decisions made in CAD shape much of what is eventually built, even modest improvements could have wide effects. We’d love to hear your feedback and ideas if you give Studio a try. We’re also hiring across product, engineering and commercial roles. If you’re interested in working on interesting problems involving AI and the physical world, please take a look at our open positions at https://cogram.com/about#careers.
2026-08-30 Zulwatha/content-parity A Measure what a website serves to machine readers. Fetches any page as several honest identities, including a cryptographically signed agent, and reports what each one received.
2026-08-23 dbl8005/sitrep-panel A A live local HTML report Claude keeps updated as it works — status board, running narrative, screenshots, newest-first log, live-reload. No cloud, no build step.
2026-11-03 SAOS-3624/dsh-see-now A 让 AI 实时看见你的屏幕：按变化量筛选帧、多帧拼进一张图，静止画面几乎不花 token
2026-10-28 onurozcan/omux A Native macOS terminal for running multiple Claude Code sessions side by side. AI coding agent manager with projects, live agent status, resumable conversations and file previews. Built with SwiftUI.
2026-10-17 Show HN: rStation – A shared research lab for your AI agents A
2026-10-05 Adolanium/hermes-gadget-sdk A Build voice and text devices for your Hermes Agent, with ESP32 firmware, a Linux client, and a desktop simulator.
2026-10-02 BootLoops-ai/bootloops A BootLoops 1.0: certified computational tools and house engines for exact and high-precision physics and quantitative science, built to be driven by LLM agents. MIT; docs CC BY 4.0.
2026-09-26 Badtheorylabs/interference-search A Reason over explicit states, many branches at once: merge duplicates, cancel dead ends, advance together.
2026-09-22 ismailperim/briefd A Your agents, briefed. Not flooded. Self-hosted context compiler: git-backed team knowledge served to coding agents as token-budgeted bundles over MCP.
2026-09-13 panda-sandeep/dearagent A Email inbox API for AI agents, self-hosted on Cloudflare Workers. Open source, with an MCP server. AgentMail alternative.
2026-09-07 alchaincyf/huashu-mac-use A 让任何 agent 操控 Mac 上没有 API 的原生 app，读后台、写不打扰、每步留取证 · Agent Skill for macOS computer use
2026-09-01 sharptoolbox/ontology-driven-dev A 本体驱动业务系统开发技能：需求探索→本体建模→应用构建三步法，基于七模型本体 YAML 与 code-paas 技术底座。支持 WorkBuddy / Claude Code / Codex。
2026-11-09 kunchenguid/headless A Agent skill that sets up a headless Mac, Linux or Windows machine over SSH, step by step, without ever seeing the admin password.
2026-11-02 pavel-molyanov/visual-review-page A Agent skill for Claude Code and Codex: shows plans and decisions as a local visual page and collects your answers
2026-10-22 xingjian-bai/opal A OPAL: a reactive optimizer. A coding agent rewrites the optimizer while your model trains.
2026-10-22 Show HN: Computer use on the phone itself: a 2B model that operates Android apps A
2026-10-20 AntoGA/agents-contract A AGENTS.md contract for AI coding agents + zero-dependency linter + SessionStart hook. Works with Claude Code, Codex, Cursor, Copilot.
2026-10-17 alonehobo/1c-cf-vendor-merge A Skill для AI-агентов: обновление доработанной конфигурации 1С на новый релиз через трёхстороннее git-слияние XML-выгрузок
2026-10-17 moritzkremb/claude-eyes A Let coding agents watch videos: index once, then skim, search, zoom, and measure editing style. Runs locally.
2026-09-30 PostHog/jeeves A Jeeves – Reasoning improves Jev-like decision models
2026-09-24 Show HN: RxFilm Studio–Create and edit your product videos with AI agent A Hi HN, I built RxFilm Studio, a video workspace that lets you create and revise videos with AI agent. The reason I started this project is simple: I’ve noticed that many independent developers and small teams want to create marketing videos for their product, but producing these videos requires a lot of time and effort — from editing and visual effects to voiceovers and subtitles. So I built Rxfilm Studio to simplify the entire video creation process. Start with an idea, a product website, images, or existing footage, and then generate a complete video with: - Story and structure - Editing and timeline - Visual and sound effects - Voiceovers, background music - Subtitles and translation Github: https://github.com/rxtech-lab/film-workflow Here's a demo marketing video created using RxFilm Studio: https://youtu.be/9_M15fRdZQE Welcome to try it out and I would love to hear your feedback!
2026-09-23 FluidInference/FluidUse A Local computer use on Apple silicon: on-device typed decisions with laya and form filling with CUA-S1-FORMS, driven through the macOS Accessibility API
2026-09-22 freestylefly/WeChatBridge A 微信聊天记录一键转发到 AI Agent 与 Obsidian 的原生 macOS 工具
2026-09-17 Show HN: Give your AI agents access to WhatsApp A Hi, I built Chat-Man because I wanted a cheap way to give my agents access to WhatsApp without integrating a WhatsApp library separately in every project. You get WhatsApp MCP server, so you can connect WhatsApp to an agent and programmatically read, search, extract and send messages - but also have a web UI for non-techies. You can also receive webhooks for incoming messages for starred conversations. What you can do with the MCP? -Read WhatsApp messages and turn them into CRM records -Summarise conversations or groups -Extract structured information from messages -Find messages, participants or other WhatsApp data -Send messages -Manage group memberships Of course.. you always need to follow data protection regulation. The web UI: -Edit group details -See who joined or left -See who is most active -Export group members -Match members against a CSV (e.g. from a CRM) -Bulk message members -Bulk kick members -Create invite links For individual contacts, you can send messages directly from the UI. Why I built it for myself? I needed WhatsApp access for several projects (hermo.ai, florahaus.co.uk, …) and didn't want to pay for a WhatsApp integration for each project. Link: https://chat-man.net I’d love to get some feedback, especially from people building agents that need access to WhatsApp. Fabian
2026-09-03 LunarXuan/image-prompt-reverse A High-fidelity AI image prompt reverse-engineering skill for Codex
2026-09-03 Mercor-Intelligence/ApexAgents-SkyRL-Recipe A RL-training knowledge-work agents on APEX-Agents data with SkyRL + Harbor
2026-09-01 Tyche-MKR/scientific-agent-skills A Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 190,000+ scientists worldwide. 165 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.
2026-08-28 christophhdesign/open-design-system-bench A How AI-ready is your design system? A plug-in benchmark that runs coding agents against your component library and grades the output. Built against a production design system and generalized for any React + TypeScript component library.
2026-08-24 w-partners/Wbrowser A Drive the Chrome you are already logged into - from your terminal or any AI assistant. Cross-platform, MCP-ready.
2026-11-05 Balanov/latch A One tiny HTTPS contract between any agent and any service. Quote, approve, commit, receipt. Zero dependencies. MIT.
2026-10-24 ethanplusai/astro-conversion-kit A Agent Skills that convert a WordPress site (or any live site) into a static Astro site with URL, SEO, content and pixel parity, then take it through launch. Works in Claude Code, Codex and any agent that reads SKILL.md.
2026-10-22 gokayfem/video-to-blender A Turn a reference video, image or idea into an editable procedural Blender node system: a coding-agent skill with a tested Geometry/Shader/compositor node-group library
2026-10-15 ananhq/anan-iphone-use A ESP32 MCP firmware for iPhone USB input, built for M5Stack StopWatch with guided board adaptation
2026-10-12 AI-BizOS/cross-harness B Open cross-harness protocol with terminal and local web interfaces
2026-10-09 beianchuqing111/agentic-kb A 法规/规程知识库智能体:混合检索(bge-m3 稠密+稀疏 → 服务端 RRF → 交叉编码器重排)+ GraphRAG + 带三道闸与审计的 ReAct 智能体。含检索质量评测设施、FastAPI 接口与 React 界面。
2026-10-09 excalidraw/cli A Official Excalidraw CLI for humans and agents
2026-10-09 finnvoor/PiDurableKit A Durable AI agents for Apple platforms: pi-durable on JavaScriptCore with a Swift API
2026-10-08 izzy-Ti/fayda-mcp A A Model Context Protocol (MCP) bridge connecting AI agents to Ethiopia's Fayda National ID (MOSIP eSignet) for zero-trust e-KYC and identity verification.
2026-10-08 Show HN: Agent.reviews – Where AI agents read and write reviews on tools B Hi HN! I’m Louis, Co-Founder of Armature (YC P26), where we help teams make their product discoverable and usable by coding agents. We already measured 50k+ agent sessions and realized that over and over agents would encounter the exact same limitations on different tasks using the same tool. So we wondered why these weren’t fixed. And the answer is simple: the feedback loop just doesn’t exist between agents and software vendors but also between different agents. Humans can share their experience on platforms like https://g2.com and https://trustpilot.com, but agents have nowhere to. So we created: https://agent.reviews: the G2 for agents. It works with a set of skills and an npm CLI (@armature-tech/agent-reviews) connecting agents to our API endpoints. Anyone can ask their agent (Claude Code, Codex, Cursor, etc.) to install it, and agents will naturally check reviews before picking a tool and post their own after using one. As usual, privacy was our main concern, so we added 3 layers before a review gets posted: Deterministic rules filtering secrets, PII, URLs, etc. A Jev classifier trained to detect any leak after the first check A small LLM checking each review to make sure nothing was missed We've been sharing this project around for a few weeks now and gathered thousands of reviews already. There are already interesting ones, for example: - A Claude Code agent noticed that the Stripe SDK systematically crashed when the API key was missing on the health check page (while it’s this page’s role to actually return an “API key missing” error) - 2 agents mentioned that Prisma required a DATABASE_URL variable even when it wasn’t connecting to any database. They both put fake URLs as a workaround, and it worked. We truly think the agent experience needs the same community effect user experience has, so everyone benefits from it: agents can pick the tools that are best optimized for them and software companies can improve their product based on real feedback. That’s why we made sure accessing reviews is free for both humans and agents and just requires copy/pasting one prompt for the agent to install our CLI & skill, start the authentication flow, and submit their first review (this helps us prevent unauthorized scraping and spam reviews). Would you let your agents submit and read reviews too? We’d love for you to set up agent reviews, ask your agent to check reviews next time it needs to pick a tool and post its own experience when using it. Then tell us how it went!
2026-10-04 Jakeschincariol/replica-skill B Eleven free Claude skills that clone any app: reverse-engineer it, rebuild it, test it for bugs, then fix what its users hate. Free, MIT.
2026-09-27 breakstageaxe61/genspark-claw B 🦀⚡ Community skill pack & CLI for Genspark Claw, OpenClaw & Hermes agents — 5 production-ready SKILL.md skills (deep research, reports, slides, browser automation, AI call prep), zero-dependency Node.js installer, MCP server, one-line install
2026-09-25 angusdevgo/Seep-Reverse-Lab B Agent-Native multi-platform reverse engineering and CWE-602 client-side authorization audit workbench.
2026-09-23 Matthew0822/MCPBastion A Zero-trust security gateway between AI agents and MCP servers: scoped capabilities, redaction, rate and cost limits, and verifiable session logs.
2026-08-28 SpatiaOS/Procedura A Agentic 3D Modeling with Procedural Control — turns a text prompt into an editable parametric program, with optional per-part materials and articulation.
2026-08-27 ooocooc/open-skill-sunset B Audit and safely retire stale generic AI agent instructions.
2026-08-23 DavidCarliez/cover A Reversible privacy proxy for AI agents: send realistic fakes, restore originals locally.
2026-08-23 kgoedecke/doop B The open-source alternative to Paper.design. A multiplayer design canvas where humans and AI agents design together, live. MCP built in.
2026-11-01 elvarofficial/teardown A A Claude Code skill that turns any machine into a 3D simulator you can take apart
2026-10-26 DenisSergeevitch/ex-ussr-kinship-research A Evidence-first genealogy skill for AI agents (Codex, Claude Code): church books, civil, military, deportation and repression records across the former USSR. Replies in your language.
2026-10-25 semsturgut/open-iphone-use A Give AI agents a real iPhone through macOS Bluetooth HID, local screen sharing, and MCP.
2026-10-21 dominikmartn/harp B Run many Claude Code and Codex agents side by side, in one sidebar that tells you which one needs you. An unofficial fork of Warp.
2026-10-21 lovisdotio/motion2code A Skills for coding agents: rebuild a motion design video as looping SVG + GSAP code, checked frame by frame against the source
2026-10-17 cranebeaker/coop-anything B Co-op Anything: point your AI agent at any single-player game and play it with friends (online and/or splitscreen!). MCP server and playbook to let you reverse engineer, fix, test, ship and share what it learns.
2026-10-16 SakanaAI/AutoIllustrate A Generate publication-ready, editable draw.io figures for research papers from a methods section, using LLM agents.
2026-10-02 Show HN: Breadcrumb, record everything on your mac + context manager for AI B Hi HN, I'm Justin. Breadcrumb records everything you do on your Mac (screen + meetings + AI transcripts + what you and your AI decided) and turns it into memory your AI can search. It's local and encrypted. You can also teach it rules by talking to it and it makes sure the right rules turn up in the right context. Works with Claude Code / Codex / Cursor / opencode. All of this is exposed to your AI as 30+ MCP tools (here's the definitions): https://innerloop.works/breadcrumb/mcp I started it in June because I wanted to understand what was going on with my AI dev work. Then it just kept growing as I jammed on it 10 hours a day for 3 months. I started by recording every screen so I could jump back to a moment and see what I'd been doing with the AI. Then I added MCP so the AI could search it. Then time tracking. Then meetings with transcripts and timestamped screenshots. After that I started feeding the data back in. At first it was just a few relevant memories in the prompt. Then I started adding rules and rules modules. Then I started recording the AI transcripts and all the other stuff the AI did (subagents, tools etc). I wanted a complete record so I could go back and see what the AI did and why. I found the more data the AI had the more useful it became. One example. I got off a zoom call with the team at the day job and said: "Review the meeting I just had, find all the bugs we discussed, and create JIRA tickets with screenshots." It read the transcript. Found every bug we mentioned, pulled the screens from the screen timeline, spun up my local server, diagnosed each bug, reviewed the code, created 14 tickets in JIRA with diagnosis and likely fixes, attached 12 screenshots as evidence, and created a nice Slack message to send to my colleagues. I never built a workflow for any of that. Claude had the meeting record, Breadcrumb tools, and rules I'd accumulated. It basically chained the steps together itself. My favorite thing about the system is observability of my own work as an AI developer. I can trace a bug all the way back to the first thing I said (or the first thing the AI did) that caused it. Then I can add a rule so it doesn't happen again. Anyway, my working system is simply Claude Desktop + Breadcrumb + Markdown. I find I don't need anything else. It's local software for one person. Local models for transcription and diarization. Data encrypted with SQLCipher and a passphrase. You can exclude any apps, websites, words and phrases. Breadcrumb itself only makes one network call which is a daily update check. (but of course your AI will use it and that is local or cloud depending on what you use). It runs cool about 12% of one core on average (M series mac has 8+). You can set it to just save OCR text or to save screenshots. If just OCR it's about 12mb day (a lot of telemetry, indexing etc). If you record screenshots on meetings it's about 120mb/hour (1 every 2 secs). It can also record screenshots of everything that's heavy at about 160mb/day. It does local classification/naming with ollama with Gemma 3 4B (2.5 GB). It's 100% free. There's no account you can just download and start using it. I'll monetize it via E2EE cloud sync and team sync. It's open beta. Just one developer working on it. Requires 16GB+ M Series Mac. Thanks for reading! https://innerloop.works/breadcrumb (homepage) https://innerloop.works/breadcrumb/download (direct download) https://innerloop.works/breadcrumb/mcp (mcp definition)
2026-10-01 Show HN: Token compression CLI to save Codex/Astra costs B Hey HN! Yolanda and Spencer here - wanted to share a token compression tool that we’ve built for ourselves to save 30% costs on codex! After maxing out sub and burning $700/day per person on api, we fine tuned a compression model to trim codex's tool call output to reduce input token + cache. It cut down tokens by 29.6% and now I just leave it on by default in Codex. To avoid messing up w/ cache, we use proxy + fine tuned qwen model trained on preserving agent trajectory to remove tool call results before they go back to the model, leaving kv cache untouched. The cli is free for everyone to use ( https://github.com/spenmcke/compress). Just lmk ur feedback and hacks to shave even more costs on astra! If you want to integrate it into your product to offer the best models at low cost, I can set you up with an sdk and api keys PS: It’s built for coding agents, not conversational agents. I optimized it for file retrieval accuracy, trajectory preservation, and quality to get up to 30% cost reduction depending on how context-heavy the task is. On security and privacy side, it's a proxy wrapping your local codex and ZDR so it doesn't retain any queries. It’s on by default in codex and when you don’t want compression, you can use `codex --uncompress` to disable it. Give it a try: code is in https://github.com/spenmcke/compress You can install the cli using `curl -fsSL https://install.everestagi.com/install.sh | sh && source ~/.config/everest/shell.sh` Love to hear any feedback and learn your hacky ways to save token costs too!
2026-09-30 openJiuwen-ai/iCode B A lightweight, extensible, fully offline development platform and agent/workflow toolkit for the AI era, with an intuitive TUI, easy deployment, and complete control over your data, agents, and workflows.
2026-09-24 palavogikos-collab/myDATA A Issue Greek invoices via AADE timologio, get the MARK, keep one JSON per invoice, sync to your ERP (SAP B1 today; SoftOne, Pylon, Megasoft wanted). Claude skill, not a ΥΠΑΗΕΣ.
2026-09-19 coldteadotai/abide B Make your coding agent abide by all your project rules
2026-09-16 Show HN: Panel – A research workspace where the agent can build its own panes B
2026-08-28 s0xDk/refactoring-ui-skill B A Claude Code skill that applies the concrete design rules from the book Refactoring UI (Wathan & Schoger) — constrained spacing/type/color/shadow scales, hierarchy through weight and color, and depth through emulated light — to building and fixing interfaces.
2026-08-26 yeswehack/claude-kit B Claude Code plugin for writing triager-grade bug bounty reports
2026-08-24 rome-os/rome B Rome is the agentic OS.
2026-08-23 yacine-baghli/DYB-Pro B DYB Pro hands the protein design pipeline to Devin: state a goal, add sequences, and an autonomous in-silico cycle returns a ranked, order-ready wet-lab shortlist at a fraction of the cost. Devin orchestrates specialist agents whose results commit to a version graph the next cycle reads.
2026-11-03 Balanov/verdict A A 197 KB local semantic verifier for AI agent tool calls. Exit code 0 is not proof.
2026-10-23 A-new/aidbg A AI-first Windows x86/x64 命令行调试器 | JSONL output, MCP server, anti-anti-debug (ScyllaHide-style), snapshot rollback, IAT rebuild, PDB types
2026-10-23 keegan0521/CinderVM A Deterministic bytecode VM for agent execution - verified images, snapshots and a hash-chained replay journal
2026-10-21 MemTensor/MemT-1.0 B Official repository for MemT-1.0
2026-10-20 C4rohan/Lampline A Self-hosted voice agent platform: import your Dialogflow CX and CX Agent Studio agents, build and test them, and take calls on your own hardware. 20 model providers, phone and SIP through jambonz.
2026-10-17 christos-antono/NPCTalk B Make any NPC in any game talk back with Fish Audio: skills, a self-hosted voice server and tools for AI coding agents, with a full GTA V example.
2026-10-15 Entelligence-AI/agentplay B Replay your coding agent sessions as a map and a timeline
2026-10-14 Zfinix/eleven A Agent-first reverse engineering for macOS, Windows and Linux: any ACP agent reads, renames and patches binaries, Android apps and GPU kernels next to you.
2026-10-11 liyupi/ai-promo-video-maker A 一句话让 AI 帮你做产品宣传片！把产品名或网址告诉 Cursor、Claude Code、Codex 等 AI 编程工具，它会自己调研产品、画出多种风格样图让你挑选、截取 4K 真实页面、按音乐节拍卡点剪辑、生成 AI 旁白并逐句转写校验、自动审图，最后交付一支 4K 商业宣传视频和可在线播放的网页版。App、网站、SaaS、开源项目、课程、AI 模型都能做，你只需确认需求、试听配乐。不用学剪辑软件，也不用花钱找外包。基于 Remotion + ElevenLabs，可直接克隆使用，也能一键安装为 Agent Skill。
2026-10-09 UniPat-AI/PaperBenchX A A benchmark for scientific research agents, evaluating end-to-end paper reproduction across 12 research areas.
2026-10-07 nickvasilescu/bops B Bops: a team of AI bots that run your business ops, each with its own computer, email and phone number. bops.bot
2026-09-27 mahaloz/ilio-mini A A tiny infrastructure for Codex to solve CTF challenges in reverse engineering. 3rd in Flare-On 13.
2026-09-24 BerriAI/liteagents A A provider-independent agent SDK with the same query() interface as the Claude Agent SDK, allowing you to use the right model for every turn.
2026-09-24 fstandhartinger/agent-message-board A Shared SQLite message board for AI coding agents, with a read-only human web UI
2026-09-18 Show HN: Aclif – Agent CLI framework: one grammar, canonical names across SaaS B Simple discovery of the provider's default schemas is possible without any credentials (since they are built into the binary). If you want it to fetch customizations of your specific instance, give it credentials to your org. The repo has more details. https://github.com/agent-cli-framework/aclif
2026-09-03 lotchuazzz-crypto/papergraph-mcp A Turn arXiv and LaTeX mathematical papers into theorem dependency graphs for AI agents through MCP.
2026-09-02 mantis5x5/kvm-automation A Agent automation for IP-KVMs: one MCP server, CLI, fleet config, and skills driving GLKVM, PiKVM, and TinyPilot. Open-source alternative to TinyPilot Automation.
2026-09-01 nickname21kmr/gameops-investigator B Evidence-first game operations incident investigation agent with read-only MCP tools and reproducible evaluations.
2026-09-01 Show HN: SlideOps – slides from a repo that flag when they drift from the code B I kept generating slide decks about my codebases with an agent, and weeks later they would go stale. Updating them was quite costly (time & tokens), as the agent would have to re-scan the whole repo to re-generate the slides. So I made the slide deck carry its own provenance: SlideOps skill turns a repo into a slide deck, with all references to files carrying exact line range and a hash. Checking is done with standard-library Python: no model calls, no network, and very fast. For example, it distinguishes MOVED (the code shifted) from CHANGED (the content differs). Updating the slides costs tokens, but now the checking part has already given the agent just the relevant context on what exactly needs to be repaired. SlideOps ships as a Claude Code plugin and runs as a plain agent skill in Codex, Copilot CLI and OpenCode. Longer write-up: https://medium.com/@lukicov/your-documentation-is-a-build-ar... GitHub repo: https://github.com/glukicov/slideops
2026-08-23 GamePhanes/GamePhanes B An open-source game coding agent environment and benchmark for Godot.
2026-11-03 CubeLink-oss/voltdir A Read-only Windows disk analysis for AI agents: one call, stable JSON, low memory.
2026-10-26 tenhkspark/secbrief A A daily vulnerability briefing for AI-agent-managed machines — deterministic matching against OSV + CISA KEV; the LLM only explains. MIT
2026-10-25 Charonyuu/whatsmyagentdoing A See what every Claude Code and Codex agent on your machine is doing — live, in a little office.
2026-10-24 actionow-ai/vvv B Build a video template once, reuse it across coding agents. Editable Remotion templates and an Agent Skill for Claude Code, Codex and 25+ agents.
2026-10-17 Tongyi-MAI/OpenGCUA A An open-source stack for generalist computer-use agents, combining code execution, GUI interaction, and cross-device control.
2026-10-11 GGGODLIN/gggodlin-cc-harness A Governance layer of a personal Claude Code harness: gates, hooks, review and spec-to-ticket workflow. Reference, not a turnkey install.
2026-10-07 bottomless/agent-duel A Blinded battles between coding agents: two agents, one prompt, you pick the winner. Desktop app, runs locally.
2026-10-06 uzairansaruzi/p3-stack B pstack reworked for T3 Code: delegated subagents, worktree threads, PR watching, scheduled runs.
2026-10-03 Show HN: Made an open-source Lego AI generator B Hi there :-) New on HN, first time posting. Past year, around December, I started experimenting with making ChatGPT and Claude generate source code in LDraw language. This LDraw is literally an "assembly" language, a low-level programming language that describes how to assemble LEGO pieces together into models, one placement instruction at a time. When executed by specific tools, like e.g. LDView, LeoCAD, Studio... these instructions become LEGO CAD models, that can be interacted with, modified, etc. Or, in other words: one LDraw source file in .mpd or .ldr format is equivalent to one LEGO CAD model. So, the idea I had was: if I manage for maybe ChatGPT or Claude to generate high-quality LDraw source files... then, they would actually be generating high-quality LEGO CAD models, right? Then, after months of iterations and trying one thing after the other... it worked!!! Long story short: using GPT-6 Astra and Opus 5.5, I've managed to create a python toolset, instructions, and docs for agents in general. Now, these can be used by them to generate LDraw models. I've packed it all as a dockerized web app for others to try and experiment, with several providers (and agents) to choose from: OpenAI, Claude and OpenRouter. Here's a bunch of exmaples: https://anteloc.github.io/index-samples.html. If you are curious about the internals of an .mpd model, the "what was the agent picturing on its mind", open the .mpd file that got your attention on a text editor, and read the first line under the ones starting with "0 FILE". I'd really appreciate feedback and comments, let's see where this goes =)
2026-09-30 silvius1996/claude-sap-kit A Claude Code plugin + ARC-1 extension + ABAP kit: run SAP reports, preview NAST print output as PDF, edit SmartForms as XML
2026-09-28 BjoernSchotte/agentglass A A looking glass for your coding agents: browse, watch, drill into and steer Claude Code, Codex & fx sessions from one tiny native TUI
2026-09-28 SecFathy/xss-specialist B KEV-gated continual-learning XSS specialist with execution-authoritative browser verification
2026-09-16 Show HN: Ordewell – turn one goal into an ordered plan of coding-agent tasks B
2026-09-13 ccompactor/ccompactor A TypeScript CLI: extract any coding-agent session (Claude Code, Codex, Pi) into a compact, verified, provenance-linked handoff another agent can continue from.
2026-09-10 Vincentwei1021/anything2explainer B Topic in, narrated explainer video out. A Claude Code / Codex skill that turns any topic into a black-canvas motion-graphics explainer video with TTS voiceover, subtitles and a chapter progress bar. Chinese or English; every frame drawn in code with Remotion.
2026-09-09 Jakeschincariol/linkedin-agent-skill B Eleven free Claude skills that run a LinkedIn account: posts off 21 hook formulas, comments, replies, profile score, weekly plan, and a humanizer that strips the AI fingerprint and scores the draft before it goes out.
2026-09-06 vinzdg/codenotch B A macOS app that pins usage limits from Claude Code, Cursor, Codex, and Antigravity to a screen edge.