Android Feed

← The 5 stories that mattered

Everything collected ·

All 377 items collected for this issue

New here? How to read this issue
  1. Read the picks. Each tab opens with 4–5 stories ranked by how much you’d regret missing them. That takes about 3 minutes. “Start the read” shows one story at a time, and “Listen” reads them aloud.
  2. Stop at “You’re caught up”, or go deeper. Everything after it is the full list of what was collected today, folded away.
  3. Android comes first. Product & AI covers what builders are launching and discussing.
Stable · Beta · Alpha
Release stages. Stable is safe to ship, beta’s API is frozen but may still have bugs, alpha is experimental.
AndroidX
Google’s Jetpack libraries (Compose, Paging, Room…), released separately from Android itself.
AOSP insider
Code changes merged into Android’s open source repository. They often show an API before it’s announced.
Lanes
Where an Android item comes from: official, AOSP insider, experts, open-source library releases, leaks and ecosystem, community.
Adopt now · Watch · Rumour
What to do about it, and how sure the source is.
Dotted underline
An inferred claim: Claude’s reading, not something the source says.

Collected automatically each morning from 37 sources. Notes are written by Claude. Links show what was trending, not endorsements.

Android5Product & AI4
Start the Android read5 stories, ranked by what you’d regret missing · 2 min read

  1. Android will stop apps writing contacts directly

    If your app saves or edits contacts, you need to move to the editor intents or apply for the new special access before the next target SDK bump.

    Google announced that apps targeting the 2027 Android release can no longer write to contacts with the WRITE_CONTACTS permission. Those writes will be rejected. Instead, apps should open the system contact editor through the standard insert and edit intents, which need no permission and are being made consistent across devices. Apps whose whole purpose is managing contacts get a new special access called MANAGE_CONTACTS, arriving in Android 17 QPR2.

    Android Developers Blog · OfficialRead the source

  2. OkHttp patches block compromised eMudhra certificates

    If you ship to set-top boxes, kiosks or other devices without Play Services, update OkHttp now.

    OkHttp, the HTTP client most Android apps use, shipped versions 5.5.1 and 4.12.1. The 5.5.1 notes say it blocks eight compromised eMudhra intermediate certificates. Phones with Google Play Services already received an automatic fix, so this update is for AOSP devices and JVM backends. The same week, Lobsters reported Indian government sites serving revoked eMudhra certificates.

    OkHttp releases · LibraryRead the source

  3. Android XR libraries reach 1.0 stable

    Android XR now has a stable first-party API, so the platform is ready for production apps rather than experiments.

    Three Jetpack libraries for Android XR went stable at version 1.0.0: SceneCore for spatial scenes, Runtime for the session layer, and ARCore for Jetpack XR for perception. Until now, building for Android XR headsets and glasses meant depending on alpha and beta artifacts that changed often. A stable base means apps can ship without chasing API churn.

    AndroidX releases · OfficialRead the source

  4. Haze 2.0 adds Glass and native backdrop support

    If you use Haze, 2.0 is a major version to plan for, and the companion post tells you whether Compose's built-in blur now covers your case.

    Chris Banes released Haze 2.0, the blur library for Jetpack Compose and Compose Multiplatform. It adds Glass, a new effects API, performance improvements and experimental support for Android's native backdrop. In a companion post he explains how Haze relates to Compose's new progressive blur, which blurs a composable's own content, and when you still need Haze.

    Chris Banes · ExpertRead the sourceProduct notes

  5. AppFunctions reaches beta on AndroidX

    The way assistants will call into Android apps is firming up, so now is the time to try exposing a function or two.

    AppFunctions, the Jetpack library that lets AI assistants call functions inside your app, published its first beta, version 1.0.0-beta01. A beta means the API shape is settling and breaking changes should be rare. Android Weekly leads this week with an explainer on the same API, and A2UI, the library for rendering agent-generated Compose UI, moved to its second alpha.

    AndroidX releases · OfficialRead the source

Also worth knowing

Releases, policy and leaks an Android engineer should know about this week.

  • 1.13.0-beta01

    Compose Multiplatform 1.13.0-beta01 shipped with a binary-incompatible change: desktop libraries using the experimental DragAndDropTransferAction must be recompiled.

    KMP teams on desktop that publish libraries using drag and drop; app teams can wait for stable.

    WatchOfficial

  • Android Studio Rabbit 2 RC 1 now available

    Android Studio Rabbit 2 reached Release Candidate 1 in the Beta channel, and Rabbit 3 opened its Canary channel.

    Teams on the Beta channel can expect Rabbit 2 stable soon; Canary users get the next cycle's features.

    WatchOfficial

Go deeper: every Android item, by lane

Official & releases

63Plan your move off WRITE_CONTACTS to the standard contact intents, note that Android XR's SceneCore, Runtime and ARCore libraries are 1.0.0 stable, and that AppFunctions is in beta while Compose 1.13 is at beta02.

AndroidX library releases

Stable · safe to ship · 8

Beta · API frozen, try it before it ships · 13

Alpha · experimental, APIs may change · 20
Show 14 more

AOSP / AndroidX insider

12Gerrit shows Remote Compose being trimmed and gaining position reporting, Health Connect adding device identifiers, and WebView preparing Google account mirroring.
Show 4 more

Experts

19Read Chris Banes on Haze 2.0 and whether you still need it now Compose has progressive blur, skydoves' three-part Compose mechanisms series, and Touchlab on hardening coroutine tests.
Show 11 more

Library releases

6Update OkHttp to 5.5.1 or 4.12.1 if you ship to AOSP devices or JVMs, since it blocks compromised eMudhra certificates; kotlinx.serialization 1.12.0 is stable, Decompose 3.6.0 is still beta.
  • Compose Unstyled 2.10.1

    Patch Changes 9658157 : Support dismissOnClickOutside for DrawerPresentation.InPlace

  • OkHttp 5.5.1new

    Version 5.5.1

  • OkHttp 4.12.1new

    Version 4.12.1

  • kotlinx.serialization 1.12.0

    This release is based on Kotlin 2.4.10 and contains all of the changes from 1.12.0-RC . Additionally, it includes a fix removing unnecessary leading whitespaces from Proguard rules ( #3270 ), and two changes stabilizing kotlin.uuid.Uuid serializer ( #3263 ) and…

  • Decompose 3.6.0-beta02

    What's Changed Library Add linuxX64 and linuxArm64 targets by @arkivanov-bot in #1028 Fix wrong directions in AndroidPredictiveBackAnimatableV1 in RTL layout by @arkivanov in #1030 Fix wrong directions in AndroidPredictiveBackAnimatableV2 in RTL layout by @arkivanov in #1031 Add…

  • Roborazzi 1.77.0new

    Bugfix: verifyRoborazzi* and compareRoborazzi* reuse the configuration cache With the configuration cache enabled, the second verifyRoborazzi* or compareRoborazzi* run after clean recalculated the task graph because "the file system entry 'build/outputs/roborazzi' has been…

Leaks & ecosystem

18OnePlus and Samsung are rolling Android 17 to phones, Google Messages gets Enhanced Protection against scams, and leaks show Pixel face unlock in the dark.
Show 10 more

Community

38Hacker News is loud about Android TV without ads, microG, a Tcl TV root exploit and Android apps on macOS, while Android Weekly leads with Compose smoothness and AppFunctions.
Show 30 more
Start the Product & AI read4 stories, ranked by what you’d regret missing · 2 min read

  1. Cloudflare open-sources a security audit skill for coding agents

    A repeatable, verified security pass is now something any team with a coding agent can run before a release.

    Cloudflare published security-audit-skill, a skill for coding agents that runs a multi-phase security audit and produces machine-readable findings that are independently verified. It was the fastest-growing repository on GitHub today, adding over two thousand stars. The idea is that an agent follows a written audit method, then a separate pass checks each finding before it is reported.

    GitHub Trending · Fastest risingRead the sourceProduct notes

  2. HTTP2Shell widens the blast radius of this week's RCE

    Check which HTTP libraries your services and apps depend on and watch for patches this week.

    The Node2Shell remote code execution issue has evolved into HTTP2Shell, and the write-up says the blast radius now covers many HTTP libraries and applications. It drew over three hundred points on Hacker News and a separate Lobsters thread. It follows the PHP2Shell bugs that Forbes reports were found with OpenAI's hacking models, and a Show HN daring people to steal Bitcoin with the exploit.

    HN front page · In 2 sourcesRead the sourceDiscussion

  3. 1Password lets agents sign in without seeing passwords

    Credential handling has been the awkward part of agent automation, and the company people already trust with passwords is now solving it.

    1Password launched 1Password for Agents on Product Hunt. It lets AI agents sign in to services on your behalf while the agent never sees the password. The vault acts as a broker, handing the agent a session instead of a secret. It is aimed at people who run browser or desktop agents that keep stopping at login screens.

    Product Hunt · New ideaRead the sourceProduct notes

  4. NVIDIA releases OpenShell, a runtime for autonomous agents

    Sandboxes for agents are becoming infrastructure, and a GPU vendor giving one away says where the market is heading.

    NVIDIA published OpenShell, described as a safe and private runtime for autonomous AI agents. It gained about nine hundred stars today and sits near twenty thousand overall. The pitch is that agents running for hours on real machines need isolation from the host and from each other, and that this layer should be open and vendor-neutral.

    GitHub Trending · Fastest risingRead the sourceProduct notes

Go deeper: themes, product notes, every source

Themes across sources

6Agent-ready Android · Agents as security auditors · Agent sandboxes · Coding-agent control surfaces · Verifying agent work · Anti-slop design skills

Ideas showing up in more than one place today. The line counts items per day over two weeks.

Agent-ready Android

Rising · 7 items from 4 sources

Google keeps wiring Android for agents. AppFunctions, the API that lets an assistant call into your app, reached its first beta on AndroidX alongside a second alpha of A2UI for agent-generated Compose UI. Android Weekly leads with an AppFunctions explainer and part six of a series on on-device agents with Google's ADK, the Developer Knowledge API serves Android docs to assistants, and Gemini 4 Argon is in Android Studio Canary.

Why now. Assistants on the phone can now plan multi-step tasks, so the missing piece is a stable, typed way for them to call app features and render app UI.

Appfunctions 1.0.0-beta01AndroidX releases

Agents as security auditors

New · 7 items from 8 sources

Security work is being handed to agents from both sides. Cloudflare's security-audit-skill, a multi-phase audit skill with verified machine-readable findings, was the fastest-growing repo on GitHub today, Argus on Show HN defends against agentic attacks, Armorer Labs launched continuous verification for agents in regulated industries, and Forbes reports OpenAI's Project Daybreak let vetted partners use hacking models to find the PHP2Shell bugs. Meanwhile HTTP2Shell shows the blast radius of one HTTP library bug.

Why now. Models can now read whole codebases and run exploits, so the same capability that creates new attack paths is the cheapest way to audit for them.

cloudflare/security-audit-skillGitHub Trending

Agent sandboxes

Rising · 5 items from 3 sources

Where an agent runs, and what it can touch, is becoming its own product. NVIDIA's OpenShell pitches a safe, private runtime for autonomous agents and gained about 900 stars today, Orgo gives agents their own cloud computers, Moli sells the fastest headless browser for agents, Papermark lets agents run a data room, and 1Password for Agents signs agents in without showing them passwords.

Why now. Agents now act on real accounts and real machines, so isolation and scoped credentials are the price of letting them run unattended.

NVIDIA/OpenShellGitHub Trending

Coding-agent control surfaces

Rising · 8 items from 3 sources

Tools for watching and steering a fleet of coding agents keep launching. Product Hunt had AgentValley in the Mac notch, Inyon for answering agents from your phone, Monish for agents on the iPhone lock screen, Peras to pause Claude Code and Codex before they burn your budget, and Bloblings as a pet per session. Agent Relay, a YC launch, lets a team's agents message each other, and claude-howto trends on GitHub.

Why now. Developers now run several long-lived agents at once, so supervising them, often away from the desk, is the new bottleneck.

InyonProduct Hunt

More on this

Verifying agent work

Steady · 7 items from 5 sources

Checking what agents produce is a product category of its own. YC launched Aucert, agents that test mobile apps, and Simulithic, AI users that test your app on every device and browser; Latent catches hallucinations before customers see them; DotMD reviews an agent's Markdown specs like a pull request; Philipp Lackner shows AI quizzes in pull requests so reviewers understand the code; and a Dev.to post asks what to do when the test fails and the model was right.

Why now. Agents write code faster than humans can read it, so the review step, not generation, now limits how much teams can ship.

Aucert - AI agents that verify mobile apps autonomouslyYC Launches

Anti-slop design skills

New · 4 items from 3 sources

Design taste is being packaged as agent instructions. Impeccable, a design language that makes your AI harness better at design, is past 86 thousand stars; Hallmark, an anti-AI-slop design skill for Claude Code, Cursor and Codex, gained about 450 today; a Dev.to author benchmarks an ai-slop-linter against humanizer tools; and Awesomic pairs AI output with human experts for the polish.

Why now. Agents can now build whole interfaces, so the generic look of their output is the most visible flaw, and a text file of taste rules is cheap to share.

pbakaus/impeccableGitHub Trending

Fading: Agent memory layers (last seen 2026-10-21) · Agent harnesses (last seen 2026-10-21) · Open-weight frontier models (last seen 2026-10-21) · Agent supply-chain safety (last seen 2026-10-21) · AI-found vulnerabilities (last seen 2026-10-21) · On-device AI (last seen 2026-10-19)

Product notes

16Who each product is for, the problem, why now, its wedge and how it makes money. Today’s picks first.

Plain text: the source says soDotted: inferred

cloudflare/security-audit-skill

repoGitHub Trending★ 2,264 today

What
A coding-agent skill from Cloudflare that runs a multi-phase security audit of a codebase and produces independently verified, machine-readable findings.
For whom
Teams using Claude Code or similar agents who want a repeatable security pass before shipping, without a dedicated security engineer.
Problem
Manual audits are slow and rare, and one-shot agent reviews produce noisy, unverified findings; teams cope with linters and occasional pentests.
Why now
Agents can now read whole repositories and run tools in phases, and a skill file is enough to encode an audit method.
Wedge
Independent verification of each finding and a machine-readable output, so results can feed CI rather than a report nobody reads.
Money
Open source, no model yet; Cloudflare gains goodwill and a path to its own platform.
Like
Semgrep · Snyk · CodeQL · OpenAI Project Daybreak

LessonPackage a process, not a prompt: the verification phase is what makes agent output trustworthy enough to automate on.

1Password for Agents

productProduct Hunt#15 on Product Hunt

What
1Password for Agents lets AI agents sign in to services on your behalf without the agent ever seeing your passwords.
For whom
People who run browser or desktop agents that need to log in to real accounts.
Problem
Agents either get pasted credentials, which leak into logs and prompts, or stop at every login screen.
Why now
Agents now browse and act on accounts routinely, so credential handling is a daily problem rather than an edge case.
Wedge
The vault people already trust becomes the broker, so the agent gets a session but never the secret.
Money
Not stated; likely part of existing 1Password subscriptions.
Like
Browser Use · Anthropic computer use · Bitwarden

LessonAn incumbent's trust asset can become the control point for a new workflow if it moves before the agent vendors do.

NVIDIA/OpenShell

repoGitHub Trending★ 896 todayHype flag

What
OpenShell, NVIDIA's open-source runtime described as a safe, private place to run autonomous AI agents.
For whom
Teams that want to run long-lived agents on their own hardware with isolation from the host and from each other.
Problem
Agents run with the developer's full permissions, so a bad tool call or prompt injection can reach real files and credentials.
Why now
Autonomous agents now run for hours unattended, which makes a sandbox the default rather than a nicety.
Wedge
A vendor-neutral runtime from a GPU company, pitched on safety and privacy rather than on a model.
Money
Open source; NVIDIA benefits when agents run on its hardware.
Like
E2B · Orgo · Docker sandboxes · Firecracker

Safe and private are claims; the listing gives no threat model or benchmark for what the runtime actually prevents.

LessonWhen you sell the hardware, giving away the runtime that keeps workloads on it is a sound strategy.

pbakaus/impeccable

repoGitHub Trending★ 619 today

What
Impeccable, a design language written for AI coding harnesses so the interfaces they generate look considered rather than generic.
For whom
Developers who let agents build front ends and have no designer to fix the result.
Problem
Agent-built UI tends to the same bland layout and spacing; people cope by hand-editing or copying a component library.
Why now
Agents now build complete apps, so visual quality is the first thing a user notices, and instructions are cheap to distribute as a repo.
Wedge
A single installable design language that plugs into existing harnesses rather than a new design tool.
Money
Open source, no model yet.
Like
Hallmark · shadcn/ui · Tailwind UI

LessonTaste can be a product when it ships in the format the tool already reads.

Nutlope/hallmark

repoGitHub Trending★ 455 today

What
Hallmark, an anti-AI-slop design skill for Claude Code, Cursor and Codex.
For whom
Developers using agent coding tools who want their generated screens to avoid the recognizable AI look.
Problem
Generated interfaces share tells such as gradients, centered hero cards and filler copy, which users now spot and distrust.
Why now
Skills became a standard way to extend several coding agents at once, so one file can reach three tools.
Wedge
Named against a specific, widely felt problem, slop, rather than a general style guide.
Money
Open source, no model yet.
Like
Impeccable · ai-slop-linter

LessonNaming the pain users already have a word for does more than describing the feature.

Moli by Lexmount

productProduct Hunt#4 on Product HuntHype flag

What
Moli by Lexmount, a headless browser built to be the fastest option for agents and harnesses.
For whom
Teams building browsing agents or scrapers that pay in latency for every page load.
Problem
General-purpose headless browsers were built for testing, not for thousands of agent page loads per minute.
Why now
Agent browsing is now a core workload, so a browser tuned for it has a market that did not exist before.
Wedge
Speed as the single pitch, aimed at harness builders rather than end users.
Money
Not stated; likely usage-based hosted browsers.
Like
Browserbase · Playwright · Puppeteer · Lightpanda

Fastest is claimed without figures or a comparison in the listing.

LessonA new workload turns an old infrastructure layer into a fresh market; speed is a fine first wedge if you can prove it.

Inyon

productProduct Hunt#39 on Product Hunt

What
Inyon lets you answer your coding agents' questions from your phone.
For whom
Developers who leave long agent runs going and do not want them stalled on a yes-or-no prompt while they are away.
Problem
Agents block on permission prompts and clarifying questions, so a run started before lunch waits until you return.
Why now
Long autonomous runs are now normal, and a phone notification is the cheapest way to keep them moving.
Wedge
Only the questions reach the phone, not the full terminal, which keeps the mobile surface small.
Money
Not stated.
Like
Monish · Claude Code remote control · Orca

LessonFind the one moment where a workflow stalls and put just that moment on the device people carry.

Aucert - AI agents that verify mobile apps autonomously

productYC Launches48 votesHype flag

What
Aucert, a YC company offering expert agents that test mobile apps autonomously and return results in under two hours.
For whom
Mobile teams shipping often who cannot keep a manual QA pass in the loop for every build.
Problem
Real-device testing is slow and costly; teams cope with thin automation suites or by skipping it before a release.
Why now
Vision-capable agents can now operate a phone screen like a tester, which was not practical two years ago.
Wedge
A time promise, under two hours, rather than a tooling pitch.
Money
Not stated; likely per-app or per-run pricing.
Like
Maestro · Simulithic · Firebase Test Lab · QA Wolf

Expert agents and actual testing are asserted; nothing shows coverage or how failures are judged.

LessonSell the outcome in the customer's unit, hours to a verdict, not the mechanism.

Simulithic — AI users that test your app on every device and browser.

productYC Launches84 votes

What
Simulithic sends AI users through your app on every device and browser to verify code as fast as agents write it.
For whom
Web and mobile teams whose agents produce more changes than their test suite covers.
Problem
Cross-device bugs are found late by real users because device matrices are expensive to test by hand.
Why now
Agent-written code has raised change volume, and agent testers can now explore a UI without scripted steps.
Wedge
Breadth of devices and browsers as the headline, pitched at teams already using coding agents.
Money
Not stated.
Like
Aucert · BrowserStack · Firebase Test Lab

LessonIf one new tool creates volume, the tool that absorbs that volume is the next product.

Agent Relay - Let your teams' coding agents message each other

productYC Launches15 votes

What
Agent Relay, real-time messaging so a team's coding agents such as Claude Code, Codex, Cursor and Conductor can message each other.
For whom
Teams where several engineers each run agents on the same codebase and keep stepping on each other.
Problem
Agents work in isolation and only learn about conflicts at merge time; humans relay context by hand.
Why now
Multiple agents per team are now common, so coordination between them is a real cost.
Wedge
Vendor-neutral, any agent on the team, which no single agent maker has reason to build.
Money
Not stated; likely per-seat SaaS.
Like
Slack · ai-memory · Conductor

LessonInteroperability between rivals is a durable wedge because none of the rivals will build it.

Show HN: Argus- open source security platform to protect against agentic attacks

productShow HN17 pts · 4 commentsHype flag

What
Argus, an open-source security platform built to protect systems against agentic attacks.
For whom
Teams that expose services to agents, or run agents, and want defenses tuned to automated attackers.
Problem
Existing security tools assume human attackers and human pace; agent-driven attacks are fast, adaptive and cheap.
Why now
The same models that write code now find and chain exploits, as the week's PHP2Shell and HTTP2Shell coverage shows.
Wedge
Open source and named around a new category, agentic attacks, rather than a general firewall.
Money
Open source, no model yet.
Like
Armorer Labs · Plango · CrowdStrike

Agentic attacks is a broad label; the Show HN does not say what the platform detects or blocks concretely.

LessonNaming a new threat category early can define a market, but the first product has to show one concrete attack it stops.

Show HN: Enzyme – a programming language that compiles wikis for agents

productShow HN20 pts · 9 comments

What
Enzyme, a programming language that compiles wikis into documentation structured for agents.
For whom
Teams maintaining internal knowledge that coding and support agents need to read reliably.
Problem
Wikis drift and contradict themselves, so agents retrieve stale or conflicting pages; people cope with hand-written context files.
Why now
Agents now consume documentation at scale, so treating docs as compiled, checkable source has a payoff.
Wedge
A compiler for docs, which can enforce structure and links the way a compiler enforces types.
Money
Not stated; likely open source with hosting later.
Like
Context Engine · llms.txt · Notion · Obsidian

LessonWhen a new reader arrives, the old document format is up for grabs.

Haze 2.0

articleChris BanesAndroid

What
Haze 2.0, the Compose Multiplatform blur library, adds Glass, a new effects API, performance improvements and experimental support for Android's native backdrop.
For whom
Android and multiplatform developers building glass-like and blurred surfaces in Jetpack Compose.
Problem
Compose had no cross-platform backdrop blur; Haze filled the gap, and now Compose's own progressive blur and Android's backdrop API overlap with it.
Why now
The platform and Compose are shipping their own blur primitives, so Haze is moving up to effects and to bridging the native backdrop.
Wedge
Already the default blur library in Compose; 2.0 keeps it relevant by wrapping the new native APIs.
Money
Open source, no model yet.
Like
Compose progressive blur · Android backdrop API · Cloudy

LessonWhen the platform absorbs your feature, move up a level and become the ergonomic layer over the platform's version.

Introducing Crayfish: An Image Cropper for Compose Multiplatform

articleDove Letter (skydoves)Android

What
Crayfish, an image cropper for Compose Multiplatform with a one-line crop dialog, a Cropper composable, aspect ratios, gestures, shapes and Coil support.
For whom
Compose and KMP developers who need profile-photo or image cropping without pulling in a View-based library.
Problem
Most Android croppers are old View libraries that do not work in Compose Multiplatform; teams write their own gesture handling.
Why now
Compose Multiplatform is now used for production apps on several platforms, so UI utilities need multiplatform versions.
Wedge
One line to get a crop dialog, plus Coil integration that most Compose apps already have.
Money
Open source, no model yet.
Like
uCrop · Android-Image-Cropper · Landscapist

LessonEvery platform shift re-opens the utility libraries; the first good Compose version of an old View library wins by default.

ExTV/Podroid

repoGitHub Trending · KotlinAndroid★ 5 today

What
Podroid, a rootless Android app that boots Alpine Linux so you can run Podman, Docker or LXC containers and desktop GUI apps on a phone.
For whom
Developers and tinkerers who want a real Linux environment on Android hardware without unlocking or rooting it.
Problem
Running containers on Android has needed root or a separate device; Termux covers shells but not container runtimes.
Why now
Phones now have desktop-class CPUs and memory, and Google's own Linux terminal preview has normalized the idea.
Wedge
Rootless, so it installs like any app and survives updates.
Money
Open source, no model yet.
Like
Termux · Android Linux Terminal · UserLAnd

LessonRemove the scariest prerequisite, root, and a niche tool becomes installable by everyone.

makecindy/cindy

repoGitHub Trending★ 477 todayHype flag

What
Cindy, an open-source AI agent pitched as working out of the box, with the slogan consider it done.
For whom
People who want a general assistant agent without assembling tools, memory and models themselves.
Problem
Most open agent frameworks need hours of setup before they do anything useful.
Why now
Enough agent building blocks exist that a bundled default configuration is now the differentiator.
Wedge
Zero-config promise and a bilingual pitch aimed at both Chinese and English users.
Money
Open source, no model yet.
Like
OpenClaw · gemini-cli · AutoGPT

Consider it done is a slogan; the listing describes no specific tasks the agent completes.

LessonIn a crowded open-source category, the setup experience is the product.

By source

11Everything collected today from GitHub Trending, Product Hunt, Show HN, HN front page, YC Launches, HF trending models, HF trending spaces, HF daily papers, Lobsters, Dev.to top, Techmeme, in each source’s own order.

GitHub Trending

10
  • cloudflare/security-audit-skillcard

    A coding-agent skill for multi-phase security audits with independently verified, machine-readable findings

    ★ 2,264 today

  • NVIDIA/OpenShellnewcard

    OpenShell is the safe, private runtime for autonomous AI agents.

    ★ 896 today

  • aws/agent-toolkit-for-awsnew

    AWS's official MCP servers and skills so agents can build on AWS; the big clouds now ship agent tooling as first-party.

    ★ 13 today

  • Nutlope/hallmarknewcard

    Anti-AI-slop design skill for Claude Code, Cursor, and Codex.

    ★ 455 today

  • luongnv89/claude-howtonew

    A visual Claude Code guide with copy-paste templates; shows how much demand there is for onboarding to agent workflows.

    ★ 392 today

  • pbakaus/impeccablenewcard

    The design language that makes your AI harness better at design.

    ★ 619 today

  • pollen-robotics/microducknew

    A tiny open-source biped duck robot with a Hugging Face simulator, a cheap entry point to legged robotics.

    ★ 319 today

  • google-gemini/gemini-clinew

    Google's open-source terminal agent for Gemini keeps trending as one of the largest agent repos.

    ★ 44 today

Show 2 more
  • jbiaojerry/ebook-treasure-chestnew

    A Chinese-language index of ebook download links; popular, but a copyright question rather than a developer tool.

    ★ 227 today

  • makecindy/cindynewcard

    Consider it done. The open-source AI agent that works out of the box · 想到，就能做到。开源、开箱即用的 AI Agent。

    ★ 477 today

Product Hunt

50
  • Llama Windowsnew

    One local AI server that every Windows app can call, the on-device model as a shared system service.

    #1 on Product Hunt

  • MRRauditnew

    Explains why your MRR numbers differ between tools and gives one you can trust, a narrow finance pain for SaaS founders.

    #2 on Product Hunt

  • Yourcast 2.0new

    AI podcast summaries with tap-to-jump timestamps inside Telegram, meeting listeners in a chat app.

    #3 on Product Hunt

  • Moli by Lexmountnewcard

    The fastest headless browser for agents and harnesses

    #4 on Product Hunt

  • Specsightnew

    Product observability for teams shipping with AI, watching what shipped rather than what was planned.

    #5 on Product Hunt

  • Papermark Virtual Data Roomnew

    Data rooms an AI agent can operate, applying agent access to a document-heavy deal workflow.

    #6 on Product Hunt

  • pawlnew

    A tiny Mac focus timer started by pulling a paw, a playful utility in a crowded category.

    #7 on Product Hunt

  • AgentGrid.ionew

    A shared workspace where humans and agents work side by side, another take on the agent control room.

    #8 on Product Hunt

Show 42 more
  • Clips by Plasticnew

    Generates lyric video ads timed to the beat for musicians, a vertical video tool.

    #9 on Product Hunt

  • How It Spreadnew

    Traces how a post went viral from the first share, analytics for creators and marketers.

    #10 on Product Hunt

  • Badgeynew

    Turns an iPhone into a name badge that faces the room, a small event utility.

    #11 on Product Hunt

  • Openstatus Incident Managementnew

    Incident management sized for small teams, a lighter alternative to enterprise on-call tools.

    #12 on Product Hunt

  • git-flow-nextnew

    A modern rework of the git-flow branching tool for teams that still use that model.

    #13 on Product Hunt

  • LinkIntelnew

    Delivers LinkedIn post analytics straight into your AI assistant.

    #14 on Product Hunt

  • 1Password for Agentsnewcard

    Let AI agents sign in for you without seeing your passwords

    #15 on Product Hunt

  • Shedinew

    Reclaims the disk space Xcode, Docker and node_modules consume, a Mac utility every developer needs eventually.

    #16 on Product Hunt

  • Byternew

    Shows exactly how much data you use on every network, a plain usage meter.

    #17 on Product Hunt

  • zCAD AInew

    AI 3D generation that outputs manufacturable parts rather than meshes, aimed at engineers.

    #18 on Product Hunt

  • AgentValleynew

    Every Claude Code agent's status in the Mac notch, a glanceable control surface for parallel agents.

    #19 on Product Hunt

  • sideOS.ainew

    Local AI for dictation, meetings and background tasks, part of the move of everyday AI onto the device.

    #20 on Product Hunt

  • River Writernew

    A quiet Mac writing app for screenplays and novels, a focused tool against bloated editors.

    #21 on Product Hunt

  • How It Wentnew

    Turns a route, photos and clips into a short film, a consumer memory app.

    #22 on Product Hunt

  • Basedash Scheduled CSV Exportsnew

    Ask once and get a fresh CSV in Slack or email every week, scheduled exports for non-technical teams.

    #23 on Product Hunt

  • Canva Craftnew

    Canva turns prompts into editable, design-ready images, keeping AI output inside its editor.

    #24 on Product Hunt

  • Pentanew

    A 1990s five-disc CD changer recreated for the iPhone Duo, a nostalgia app built for the new foldable.

    #25 on Product Hunt

  • Many Displaysnew

    Restores Mac windows after sleep or replugging displays with zero clicks, a long-standing annoyance fixed.

    #26 on Product Hunt

  • myPipnew

    Vibe-code, publish and monetize games with AI, a creator pipeline for casual games.

    #27 on Product Hunt

  • Context Enginenew

    One shared memory for AI agents and human teammates, memory as a team resource rather than per tool.

    #28 on Product Hunt

  • MOVO AInew

    A long-form video agent for mobile, generating full videos from the phone.

    #29 on Product Hunt

  • Product Hunt Visual Launch Editornew

    Product Hunt's own visual editor shows your launch page as you build it.

    #30 on Product Hunt

  • Agent Huntnew

    Live rankings of which tools AI agents use most, a leaderboard for the agent tool market.

    #31 on Product Hunt

  • Ovren Live Previewnew

    See a product change live before it ships, preview environments for product teams.

    #32 on Product Hunt

  • Polylane for Cloudflarenew

    An AI on-call engineer for Cloudflare apps, vertical agent ops for one platform.

    #33 on Product Hunt

  • Dotsnew

    A customizable workspace in the Mac notch, the notch as a new UI surface.

    #34 on Product Hunt

  • ElevenLabs Realtimenew

    ElevenLabs' realtime voice agent models with 570 milliseconds to first audio, pushing voice agents toward conversational latency.

    #35 on Product Hunt

  • FLUX 3 Maxnew

    FLUX 3 Max on fal generates ten seconds of video in about twelve seconds, near real-time video generation.

    #36 on Product Hunt

  • Instant Searchnew

    Fast local AI search over your files, keeping file search on the device.

    #37 on Product Hunt

  • ShogunAInew

    A personal AGI on your PC, a big claim with no stated wedge.

    #38 on Product Hunt

  • Inyonnewcard

    Answer your coding agents from your phone

    #39 on Product Hunt

  • Monishnew

    Your AI agents live on the iPhone lock screen, agent status as a glanceable widget.

    #40 on Product Hunt

  • Cloudflare Clairicnew

    Cloudflare's own model for fast, cheap data extraction, infrastructure companies shipping task-specific models.

    #41 on Product Hunt

  • Headroomnew

    Your Mac on the iPhone Duo; fold it and a keyboard appears, an early app built around the new foldable.

    #42 on Product Hunt

  • Gertynew

    A private ChatGPT alternative that sets itself up on your Mac, local assistants for non-technical users.

    #43 on Product Hunt

  • Dub Startup Programnew

    Dub's startup program for partner-led growth, a go-to-market offer rather than a product.

    #44 on Product Hunt

  • Bloblingsnew

    A pet for every Claude Code session, a playful way to see which agents are alive.

    #45 on Product Hunt

  • Dots Phonenew

    A phone-calling plugin for OpenAI's Dots, giving an assistant the ability to make calls.

    #46 on Product Hunt

  • DotMDnew

    Review your agent's Markdown specs like a pull request, treating plans as reviewable code.

    #47 on Product Hunt

  • Trufflenew

    Fixes messy tags, covers and duplicates in a Mac music library, a classic utility.

    #48 on Product Hunt

  • Perasnew

    Pauses Claude Code and Codex before they burn your weekly budget, a spend guard for agent subscriptions.

    #49 on Product Hunt

  • Plangonew

    Use AI to defend against AI, a security pitch with no concrete mechanism in the listing.

    #50 on Product Hunt

Show HN

6

HN front page

30
Show 22 more

YC Launches

20
Show 12 more

HF trending models

18
Show 10 more

HF trending spaces

16
Show 8 more

HF daily papers

20
Show 12 more

Lobsters

25
Show 17 more

Dev.to top

20
Show 12 more

Techmeme

15
Show 7 more

About this issue

377 items from 77 sources, 12 of them seen in several sources. Every source was fetched.

Source health

77Which sources were fetched today, and how many items each returned.
SourceSectionItemsStatus
GitHub TrendingMain10Fetched
Product HuntMain50Fetched
Show HNMain6Fetched
HN front pageMain30Fetched
YC LaunchesMain20Fetched
HF trending modelsMain20Fetched
HF trending spacesMain20Fetched
HF daily papersMain20Fetched
LobstersMain25Fetched
Dev.to topMain20Fetched
TechmemeMain15Fetched
AndroidX releasesAndroid · Official & releases45Fetched
Android Developers BlogAndroid · Official & releases5Fetched
Android Studio releasesAndroid · Official & releases3Fetched
Kotlin blogAndroid · Official & releases3Fetched
Kotlin releasesAndroid · Official & releases0Fetched
Compose Multiplatform releasesAndroid · Official & releases4Fetched
Android Developers YouTubeAndroid · Official & releases3Fetched
Kotlin on YouTube (JetBrains)Android · Official & releases3Fetched
Google Developers BlogAndroid · Official & releases1Fetched
AndroidX merged API changesAndroid · AOSP / AndroidX insider12Fetched
Dove Letter (skydoves)Android · Experts5Fetched
skydoves on MediumAndroid · Experts0Fetched
Jake WhartonAndroid · Experts0Fetched
Zac SweersAndroid · Experts0Fetched
Chris BanesAndroid · Experts3Fetched
Ryan HarterAndroid · Experts0Fetched
Talking KotlinAndroid · Experts0Fetched
Joe BirchAndroid · Experts2Fetched
CommonsWare (Mark Murphy)Android · Experts3Fetched
TechYourChance (Vasiliy Zukanov)Android · Experts1Fetched
Kt. Academy (Marcin Moskała)Android · Experts2Fetched
Touchlab (Kotlin Multiplatform)Android · Experts3Fetched
Landscapist releasesAndroid · Library releases0Fetched
Balloon releasesAndroid · Library releases0Fetched
Sandwich releasesAndroid · Library releases0Fetched
Cloudy releasesAndroid · Library releases0Fetched
Compose Stability Analyzer releasesAndroid · Library releases0Fetched
Haze releasesAndroid · Library releases0Fetched
Compose Unstyled releasesAndroid · Library releases1Fetched
Coil releasesAndroid · Library releases0Fetched
Circuit releasesAndroid · Library releases0Fetched
Metro releasesAndroid · Library releases0Fetched
Molecule releasesAndroid · Library releases0Fetched
Turbine releasesAndroid · Library releases0Fetched
Paparazzi releasesAndroid · Library releases0Fetched
SQLDelight releasesAndroid · Library releases0Fetched
OkHttp releasesAndroid · Library releases2Fetched
Retrofit releasesAndroid · Library releases0Fetched
LeakCanary releasesAndroid · Library releases0Fetched
Koin releasesAndroid · Library releases0Fetched
Dagger / Hilt releasesAndroid · Library releases0Fetched
KSP releasesAndroid · Library releases0Fetched
kotlinx.coroutines releasesAndroid · Library releases0Fetched
kotlinx.serialization releasesAndroid · Library releases1Fetched
Ktor releasesAndroid · Library releases0Fetched
Apollo Kotlin releasesAndroid · Library releases0Fetched
Kermit releasesAndroid · Library releases0Fetched
Store releasesAndroid · Library releases0Fetched
Decompose releasesAndroid · Library releases1Fetched
Orbit MVI releasesAndroid · Library releases0Fetched
Roborazzi releasesAndroid · Library releases1Fetched
detekt releasesAndroid · Library releases0Fetched
ktlint releasesAndroid · Library releases0Fetched
MockK releasesAndroid · Library releases0Fetched
Lottie releasesAndroid · Library releases0Fetched
Authority Insights (leaks & teardowns)Android · Leaks & ecosystem3Fetched
9to5Google · AndroidAndroid · Leaks & ecosystem8Fetched
Android AuthorityAndroid · Leaks & ecosystem6Fetched
Android FaithfulAndroid · Leaks & ecosystem1Fetched
Android WeeklyAndroid · Community15Fetched
GitHub Trending · KotlinAndroid · Community10Fetched
Lobsters · androidAndroid · Community2Fetched
HN · AndroidAndroid · Community11Fetched
ProAndroidDevAndroid · Community0Fetched
Kotlin WeeklyAndroid · Community1Fetched
Philipp Lackner (YouTube)Android · Community2Fetched
1 of 5

Android will stop apps writing contacts directly

If your app saves or edits contacts, you need to move to the editor intents or apply for the new special access before the next target SDK bump.

Google announced that apps targeting the 2027 Android release can no longer write to contacts with the WRITE_CONTACTS permission. Those writes will be rejected. Instead, apps should open the system contact editor through the standard insert and edit intents, which need no permission and are being made consistent across devices. Apps whose whole purpose is managing contacts get a new special access called MANAGE_CONTACTS, arriving in Android 17 QPR2.

Android Developers Blog · OfficialRead the source

1 of 4

Cloudflare open-sources a security audit skill for coding agents

A repeatable, verified security pass is now something any team with a coding agent can run before a release.

Cloudflare published security-audit-skill, a skill for coding agents that runs a multi-phase security audit and produces machine-readable findings that are independently verified. It was the fastest-growing repository on GitHub today, adding over two thousand stars. The idea is that an agent follows a written audit method, then a separate pass checks each finding before it is reported.

GitHub Trending · Fastest risingRead the source