DEV Community: AIFix

lovable security: A 3-Step Supabase Check After a Study of 159,041 Lovable Apps

AIFix — Sat, 14 Nov 2026 19:33:07 +0000

Lovable security is getting attention this weekend. Security firm Crissy Field, working with Handelsblatt, looked at 159,041 Lovable projects and found many apps where personal data could be read without logging in. If you built a client site, a booking form or an internal tool with Lovable (or any AI app builder) on top of Supabase, here is what the study found and a short check you can run today. This is a research-based summary, not a hands-on audit.

What the study found

From Crissy Field's research report:

They found 159,041 Lovable projects in Common Crawl data from December 2025 to September 2026. 34,545 of them used Supabase.
24,664 sites sent Supabase records as soon as the page loaded, with no login. That alone isn't a problem, since some data is meant to be public.
With further read-only requests using the same public key, they found tables with email addresses on 7,403 of those 24,664 sites (30%). The authors say this is not a confirmed leak rate, and their sample doesn't represent every Lovable app.
Examples included forum profiles and chat messages, ticket and food orders, identity document details, contracts and invoices.

They also point out that Lovable's terms put responsibility for apps and data handling on the customer. The study was discussed widely on Hacker News.

Why this happens

A Supabase front end talks to the database with a public key. What keeps data private is Row Level Security (RLS) plus the right grants. The Supabase RLS docs are blunt: a table in an exposed schema without RLS is readable and writable by any role with a grant on it. On many projects, anon (signed-out visitors) and authenticated get every privilege on new public tables by default, and adding policies doesn't take those grants back.

A 3-step lovable security check

1. Turn on RLS for every exposed table. In the Supabase dashboard, check each table in public. Any table without RLS is a red flag.

alter table public.orders enable row level security;

2. Write a policy for each table, and trim the grants. A typical rule lets people read only their own rows:

create policy "Users read own orders"
on public.orders for select
to authenticated
using ( (select auth.uid()) = user_id );

-- signed-out visitors shouldn't touch this table at all
revoke all on public.orders from anon;

Give anon only what visitors really need to see, like a public menu or price list.

3. Test it logged out, and keep secrets out of the browser. Open the app in a private window without signing in and see what loads. The Supabase docs also say never to put a secret (service_role) key in the browser, because it bypasses RLS. Repeat this check after every big change the AI makes to your app.

When to call in help

Crissy Field's advice is simple: an easy builder doesn't remove the need for a proper review of access permissions before an app handles real customer data, and automated checks don't replace a professional assessment. If your app stores customer emails, payments or ID details, get a second pair of eyes on it.

More research-based AI news for small businesses and creators at ainews.useaifix.com. Need help with AI automation for your business? See useaifix.com.

Agentic Payments for Small Teams: What x402 and ChatGPT Wallet Mean for Your Automations

AIFix — Fri, 13 Nov 2026 19:28:15 +0000

Agentic payments had a busy week. On 12 November, Brave let AI agents pay for search results one request at a time using the x402 protocol, and on 13 November OpenAI launched ChatGPT Wallet so its agents can complete purchases. If you build automations for a small business, in n8n, Make, or your own scripts, this is worth understanding now, before an agent with a budget ends up in one of your workflows. This is a research-based overview, not a hands-on review.

Two models of agentic payments

1. Pay-per-request (x402). x402 describes itself as an open, neutral standard for internet-native payments between clients and servers, aimed at uses like API monetization and agentic commerce. The idea reuses the old HTTP status code 402 Payment Required.

Brave's announcement shows how it works in practice:

The agent calls a Search API endpoint with no API key.
The server replies 402 Payment Required with the price and payment details.
The agent signs a USDC payment for that amount on the Base blockchain and retries.
Once the payment is verified, the search runs and the result comes back with a receipt.

Brave prices these calls at $0.005 each ($5 per 1,000), says no Brave Search API account is needed, and points to Purl, a CLI wallet developed by Stripe, as the easiest way to start.

2. Human-approved card payments (ChatGPT Wallet). OpenAI's ChatGPT Wallet takes the opposite route. A person approves every purchase with Face ID or a fingerprint, and the agent gets a payment token instead of the real card number. It starts with US ChatGPT Pro subscribers inside dots.

The difference matters for builders. With x402, the signed payment is the credential: whoever holds a funded wallet can spend. With the card model, a human stays in the loop for each purchase.

What this means for small teams

Fewer API keys to manage. For pay-per-call services, a wallet can replace sign-ups and keys, which is handy for one-off scripts and prototypes.
Costs move closer to the work. You pay per request instead of per plan, which suits low-volume automations.
Your budget controls move into your code. With an API plan, the vendor's dashboard often caps spending. With per-request payments, your setup has to do that job.

Safe setup tips

If you experiment with agentic payments in an automation, these basics go a long way:

Use a separate wallet or card for agents. Never connect your main business account. Fund it with only what you're prepared to lose.
Set hard spending limits. Put a per-call cap and a daily cap in your own code, not just in a prompt. A retry loop can fire thousands of valid small payments.
Require approval above a threshold. Small, known calls can run alone; anything new or larger should wait for a human "yes" (a Slack or email approval step works fine).
Allowlist where money can go. Only let the agent pay endpoints or merchants you've approved.
Log every payment. Record the time, endpoint, amount, receipt and the task that triggered it, and review it weekly.
Keep wallets away from untrusted input. An agent that reads emails or web pages shouldn't hold an unlocked wallet in the same step.
Start on a testnet if the service offers one, before using real funds.

If you sell APIs or products

OpenAI warns that some sites and fraud systems may block legitimate agent purchases, and it published a Commerce Partner Playbook with steps such as letting signed agent requests through your CDN or firewall. If you charge for an API, x402-style pricing is also an option to keep in mind for low-volume callers.

Bottom line

Agentic payments are early, and the two big approaches, pay-per-request and human-approved cards, carry very different risks. Start small, cap spending in code, and log everything.

More research-based AI news is at ainews.useaifix.com, and if you'd rather have safe automations built for you, see useaifix.com.

gemini 4 hydrogen: What Automation Builders Should Know About Google's Fast New Model

AIFix — Thu, 12 Nov 2026 19:30:02 +0000

The gemini 4 hydrogen model is Google's new fast option for everyday coding and agent work, launched on 12 November 2026. If you build automations for a small business, such as n8n or Make workflows, support bots or little internal tools, this is the Gemini 4 model most likely to end up doing the day-to-day work. Here's a research-based summary of what Google says it does and how to think about using it.

What it is

Google now has three Gemini 4 models. According to Google's launch post, Gemini 4 Carbon and Gemini 4 Argon are the frontier models for deep, long-horizon reasoning, while Hydrogen is "built for rapid execution across diverse agentic workflows." In plain words: the big models plan and untangle hard problems, and Hydrogen does the fast, repeatable steps.

Google says it's available across all tiers of Google AI Studio and the Gemini API, plus Gemini for business, the Gemini app, Google Search, and Google Antigravity for paid AI Pro and Ultra subscribers.

What Google claims it improves

These are Google's own figures, so treat them as a starting point and test on your own tasks:

It finishes tasks 2.5x faster than Gemini 3.8 Flash at its high default setting.
Terminal-Bench 4.1 goes from 18.2% (3.8 Flash) to 57%.
It cuts fabricated claims by more than 5x and "adapts cleanly when a tool call fails or a search returns empty."
Native 1 million-token input with up to 1 million output tokens, across PDFs, spreadsheets, charts, audio and video.

The developer post adds that it runs across four configurable thinking levels, and 9to5Google reports that Antigravity labels them Low, Medium, High and Extra.

Why automation builders should care

The most useful part for small teams is the multi-model pattern. Google describes a setup where Carbon acts as the coordinator and Hydrogen powers "worker and verifier agents." It also launched Orchestrator in Antigravity, which switches between Carbon and Hydrogen in one session. On Terminal-Bench 4.0, Google says Orchestrator reached near-Carbon success rates while using 24% less quota and finishing 19% faster on average.

You can copy the same idea in your own workflows without Antigravity:

Route by difficulty. Send classification, extraction, summaries and simple replies to the fast model. Send planning or tricky edge cases to a bigger model.
Add a verifier step. Have a second, cheap call check the first output (does the JSON match the schema, does the reply answer the question) before anything reaches a customer.
Handle empty results on purpose. Google highlights better behaviour when tools fail. Still, give your workflow a clear fallback branch when a search or API call returns nothing.
Keep long documents in one call. With a 1M-token input window, you can often pass a full PDF or spreadsheet instead of chunking it first.

Before you switch

Price. Don't assume it costs the same as Flash. Check the current Gemini API pricing page before swapping a high-volume workflow.
Test with your own data. Run 20 to 50 real examples through your current model and Hydrogen, and compare accuracy, speed and cost side by side.
Check the model card. Google DeepMind's model card lists intended uses, including sub-agent work and iterative coding.

Bottom line

Gemini 4 Hydrogen looks aimed squarely at the boring, high-volume steps that make up most real automations. If you already use Gemini 3.8 Flash in a workflow, it's worth a side-by-side test this week.

More research-based AI news for small businesses and creators is at ainews.useaifix.com, and if you'd rather have these workflows built for you, see useaifix.com.

gemini decisions api: How Small Teams Can Use It to Sort Leads and Tickets

AIFix — Wed, 11 Nov 2026 22:27:41 +0000

The gemini decisions api is Google's new endpoint for one narrow job: making fast, bounded choices. Google launched it on 11 November 2026 with a new model, Gemini Decision 1 (gemini-decision-1). If you run automations for a small business, such as sorting support tickets, scoring leads or checking order photos, this is worth a look. Here's a research-based summary of what it does and how you might use it.

What it is

Most AI models write an answer word by word, and then your code has to parse it. Gemini Decision 1 doesn't write text at all. According to Google's announcement, it scores your candidate options in a single pass and returns probabilities and confidence scores.

You send one shared input (text, images, audio, video or PDFs, up to 256K tokens) plus up to 1,000 independent questions. There are three question types:

noul: a yes/no check that returns the probability the condition holds.
choice: picks one of 2 to 255 options and returns a probability for each.
score: grades the input against an ordered rubric.

Every answer includes a confidence score from 0 to 1, so your workflow can send unsure cases to a fallback or a person.

What it costs and how fast it is

Google lists it at $0.10 per 1M input tokens, with $0 for output tokens. The shared input is billed once per request, however many questions you ask. There's a no-cost tier in Google AI Studio for prototyping. Google reports 150-250 ms end-to-end latency for inputs up to 16K tokens.

MarkTechPost's write-up adds useful caveats: it's Google-hosted only (v1beta, google-genai SDK 3.2 or later), there's no prompt caching or batch option yet, and the benchmark figures are self-reported by Google.

A simple example: triaging a support message

Here's a sketch based on the request shape in Google's post. One message, three questions, one call:

from google import genai

client = genai.Client(http_options={"api_version": "v1beta"})

result = client.decisions.create(
    model="gemini-decision-1",
    input=[{"type": "text",
            "text": "I was charged twice for my order. Please refund one."}],
    questions={
        "wants_refund": {
            "type": "noul",
            "instructions": "Is the customer asking for a refund?",
        },
        "team": {
            "type": "choice",
            "instructions": "Which team should handle this?",
            "criteria": {
                "billing": "Payments, refunds, invoices.",
                "shipping": "Delivery, tracking, damaged parcels.",
                "general": "Anything else.",
            },
        },
        "urgency": {
            "type": "score",
            "instructions": "How urgent is this message?",
            "criteria": ["Not urgent.", "Needs a reply today.", "Needs a reply now."],
        },
    },
)

Check the official docs for the exact response fields before you wire it into production.

Where it fits in a small-business workflow

Lead scoring: ask "Is this a real buyer?" and "Which package fits?" on every form submission.
Ticket routing: pick the right team and urgency before a person opens the inbox.
Order checks: pair a customer photo with the order note and ask "Does the item look damaged?"
Agent guardrails: a quick yes/no before an AI agent runs a risky tool.

A practical rule: act on high-confidence answers automatically, and send low-confidence ones to a human queue. That keeps the speed without trusting every guess.

Google says integrations are rolling out across OpenRouter, LangChain, Vercel, LiteLLM and others, so you may be able to call it from tools you already use.

For more research-based AI news for small businesses and creators, visit ainews.useaifix.com. If you'd like help setting up automations like this, see useaifix.com.

claude data: Setup Checklist for Connectors, Certified Queries and Live Sharing

AIFix — Tue, 10 Nov 2026 19:24:01 +0000

claude data is now generally available. On 10 November 2026, Anthropic renamed Claude Dashboards (in beta since October) to Claude Data and opened it to all paid plans (Anthropic). If you are the person on a small team who wires up tools, here is a practical, research-based checklist from the announcement.

1. You need a data connector first

Anthropic says Claude Data needs a data connector, such as a warehouse or CRM. The announcement names Amazon Redshift, BigQuery and ClickHouse as sources teams already query (Anthropic). No connector, no dashboards. Start by deciding which single source holds the numbers you actually report on.

2. Check who can turn it on

Team plans: on by default.
Enterprise plans: Claude Data turns on by default on 19 November 2026.
Admins can change this in Organization settings > Artifacts (Anthropic).

If you are on Enterprise and not ready, that date is your deadline to decide.

3. Use certified queries to stop metric drift

Certified queries (in beta) let your data person approve the queries everyone should use. Charts built on them carry a "certified" badge (Anthropic). For a small team, certify only the handful of metrics that matter, such as monthly revenue, new customers and churn, so nobody builds a dashboard on a home-made definition.

4. Understand live sharing before you send links

Live sharing (also beta) lets colleagues open a shared dashboard and see current numbers even if they lack access to the connectors. Viewers see results, not the underlying data, and can run only the queries you approved (Anthropic). That is handy, but treat a shared link like a report: share only what that person should see.

5. Set refresh cadence per dashboard

Dashboards now support caching, and you can set refresh cadences per dashboard. Previously every dashboard reloaded on every load (Anthropic). Slower-moving numbers can refresh less often, which keeps things fast.

6. Turn on your review habits

Claude Data adds tracked changes, so you can see who edited what and roll back to a version. It can also draft data stories or narrated videos from your metrics, which Anthropic says are for you to review before sharing (Anthropic; iClarified). Keep a human check on any narrative before it reaches a client.

Quick setup order

Pick one source of truth and connect it.
Certify 3 to 5 core metrics.
Build one dashboard on certified metrics only.
Set its refresh cadence.
Share with one colleague and confirm what they can and cannot see.

Partners listed for Claude Data include Amplitude, Mixpanel, Omni, PostHog and Sigma, if your product analytics already live there (Anthropic).

More research-based AI tool breakdowns for small teams at ainews.useaifix.com, and automation help at useaifix.com.

claude fable 5.2: API Price, Claude Code Setup, and How It Differs from Mythos 5.2

AIFix — Tue, 10 Nov 2026 09:01:40 +0000

The new model IDs, API prices, Claude Code switch command and the 30-day data-retention rule, from official docs.

If you use the Claude API or Claude Code, the new model ID is claude-fable-5-2. Anthropic released Claude Fable 5.2 on 9 November 2026, alongside Claude Mythos 5.2 (claude-mythos-5-2). The docs describe them as two configurations of the same new model. Fable 5.2 is generally available. Mythos 5.2 is limited to organisations that pass Anthropic's Cyber or Life Sciences Verification Programs.

API pricing for both is $8 per million input tokens and $40 per million output tokens (USD), down from $10 and $50 on the 5.1 versions. Fast mode is listed at $16 and $80. The context window is 1 million tokens, with up to 128,000 output tokens per request.

To try Fable 5.2 in Claude Code, the docs say you need version 2.1.321 or later, then run /model fable or start with claude --model claude-fable-5-2. No Fable model is the default, so you have to choose it. Anthropic also says Fable models use plan limits faster than other Claude models. One other practical note: both 5.2 models require 30-day data retention and are not available under zero data retention unless Anthropic expressly authorises it.

Anthropic still suggests Opus 5.5 as the daily driver, at $4 input and $20 output, and moving up to Fable for harder, open-ended work. Benchmark figures in the coverage are Anthropic's or Artificial Analysis's, not our tests. This is a research-based summary; we have not tested the models.

Prices, access rules and the official benchmark table are in the full article: https://ainews.useaifix.com/news/claude-fable-5-2-mythos-5-2-launch/

crowdstrike ai kill switch: How Small Teams Can Build Their Own Off Switch for AI Agents

AIFix — Wed, 04 Nov 2026 04:25:17 +0000

The crowdstrike ai kill switch landed on 3 November, and it's one of the most talked-about AI security stories this week. CrowdStrike says it lets a security team stop an AI agent that goes off-script, block it from launching again and cut its network access. Most small teams won't buy an enterprise security platform, but the idea behind it is worth copying with the tools you already have.

What CrowdStrike actually announced

According to CrowdStrike's press release, the AI Kill Switch is part of Falcon Guardian and does four things:

stops the running agent, first at CrowdStrike's AI Gateway and "soon" on the endpoint itself,
blocks it from relaunching on the targeted host or host group,
cuts off its enterprise network access,
and logs every action, with an authorized admin pulling the trigger.

The company's investor release adds agent-level hooks that enforce policy inside tools such as Claude Code, OpenAI Codex, GitHub Copilot and Cursor. CNBC's coverage frames it as a tool for internal agents that go rogue.

One honest catch: today the stop happens at the gateway. Endpoint-level termination is listed as coming soon, so an agent running a local model that never touches the gateway isn't covered by that part yet.

A small-team version you can set up today

You don't need an enterprise contract to get most of the benefit. The pattern is the same: one place to say no, a way to stop things fast, and a record of what happened.

1. Put a gate in front of risky tool calls. If you use Claude Code, its hooks let you run a script before any tool call. A PreToolUse hook can return deny, ask or allow, or exit with code 2 to block the call. The hooks guide shows a script that blocks any shell command containing drop table:

#!/bin/bash
INPUT=$(cat)
COMMAND=$(echo "$INPUT" | jq -r '.tool_input.command')
if echo "$COMMAND" | grep -q "drop table"; then
  echo "Blocked: dropping tables is not allowed" >&2
  exit 2
fi
exit 0

Extend the same idea to rm -rf, git push --force, or anything that touches production.

2. Make "ask" the default for writes. Let agents read freely, but send any write, deploy or payment action back to a human. That's the human-approval step CrowdStrike puts at the centre.

3. Have a real off switch. For most small setups, the fastest kill switch is the API key. Give each agent its own key so you can revoke one without breaking everything else, and set a spending limit with your model provider.

4. Check your agent config now and then. This week also showed that other apps can edit your agent settings. A Postman bug report covered by The New Stack says the desktop app installed plugins into Claude Code and Codex without clear consent. Look at your plugin, hook and MCP server lists once a week.

5. Keep a log. Even a simple file of what each agent ran, and when, makes it much easier to see what went wrong.

The takeaway

The crowdstrike ai kill switch is an enterprise product, but the lesson is simple: an agent you can't stop quickly isn't ready for real work. A blocking hook, an "ask before writing" rule and one revocable key per agent get a small team most of the way there.

This is a research-based summary of public sources. More plain-English AI news for small businesses and creators is at ainews.useaifix.com, and AI automation help is at useaifix.com.

deepseek creator mode: What It Is, How to Try It, and How to Stay Safe

AIFix — Mon, 02 Nov 2026 19:25:49 +0000

deepseek creator mode is a new way to build add-ons for DeepSeek Harness by simply describing what you want. DeepSeek introduced it on Monday, 2 November 2026, and Bloomberg reported that the Harness team says it lets users "turn ideas into plugins that you can use." This research-based guide covers what it is, the official setup steps, and how small teams can use it safely.

What DeepSeek Harness and Creator Mode are

DeepSeek Harness (dsh) is DeepSeek's open-source agent app, released under the MIT licence and built on an "everything is a plugin" design. On its official page, DeepSeek says it is in public preview worldwide and can help with everyday work (documents, data, slides), coding, research and scheduled background tasks. You can run it as a desktop app or launch a web interface. On 3 November, the GitHub API showed the repo with more than 269,000 stars.

Creator mode is the part that extends the app itself. According to that page, you can "install plugins, or create them through chat in Creator mode" to add tools, skills and interface pieces. The page's demo asks for a Pomodoro timer plugin, and the agent writes, installs and checks it. Quartz lists a design workspace, a sandbox game and a 3D model visualiser as examples DeepSeek shared. Investing.com adds a 3D model rendering plugin and a video editing workbench, and notes that the agent is aimed at developers rather than casual users.

How to install and try it

These steps come straight from the project README and Web UI guide:

Install Node.js.
Start the web interface:

   npx @deepseek-ai/dsh web

It starts at http://127.0.0.1:3080 and opens your browser.

Add a model. Go to Settings → Models and paste a DeepSeek API key. The model guide explains how to add other providers or an OpenAI-compatible endpoint.
Pick a workspace with Choose workspace. You can't send messages until one is selected.
Switch to Creator mode and describe a small plugin in plain words, for example a simple timer or a checklist panel.

If you prefer the source, the README also lists git clone, then pnpm install, pnpm run build and pnpm dsh web.

Keep in mind that the app is free and open source, but the model calls you make with your API key may be billed by the provider.

Safety tips for small teams

DeepSeek is unusually direct about the risks. Its safety notice says Harness has not had a security audit, can run code and commands the model writes, and can load third-party plugins that may change or delete files or leak data or credentials. The project also warns of compatibility-breaking changes during the developer preview. A design note in the repo explains that installing plugins needs full access or a one-time approval, because plugin code runs with the app's own permissions and stays installed across sessions and restarts.

What that means in practice:

Read the plugin before you approve it. Creator mode writes real code. Open the generated files and make sure they only do what you asked.
Keep approval prompts on. Stick with the default "workspace write, ask first" style of permission rather than full access, so you see each sensitive step.
Run it somewhere disposable. DeepSeek suggests a throwaway virtual machine, container or separate environment, plus backups of any files it can reach.
Protect your API keys. Keys are stored locally (in $DSH_HOME/.credentials.yaml, per the docs). Don't put client passwords or payment logins in the workspace, and use a separate, spend-limited key if your provider supports one.
Only add plugin sources you trust. The plugin index guide calls indexes third-party content and recommends https: links.

Who it suits

Creators and small businesses comfortable with a terminal can use it to build small helpers without writing code. For client-facing work, treat it as an experiment for now and keep a person in charge of every install.

For more research-based AI news, visit ainews.useaifix.com. For help with AI automations for your business, see useaifix.com.

paperclip ai: How to Set Up the Free, Open-Source App for Managing AI Agents

AIFix — Mon, 02 Nov 2026 13:23:46 +0000

paperclip ai is trending again on GitHub this week, and a lot of people are searching for a "paperclip ai tutorial" or "paperclip ai setup". If you run a small business or a small dev team and you already use one or more AI coding agents, here's a plain, research-based walkthrough of what Paperclip is, how to install it, and what to watch before you let agents run on their own.

What Paperclip is

According to its GitHub README, Paperclip is "open-source orchestration for teams of AI agents". In practice it's a Node.js server with a React dashboard where you:

set a goal for the work,
"hire" agents into roles on an org chart ("any bot, any provider", in the README's words),
assign tasks, approve plans, set budgets, and watch progress from one place.

It's MIT-licensed and self-hosted, and the README says no Paperclip account is needed. The repo had just over 110,000 stars when we checked the GitHub API today.

It works with agents you may already use. The README lists OpenClaw, Claude Code, Codex, Cursor, Gemini CLI, OpenCode, Pi, Hermes and Kimi Code, among others, plus custom processes and HTTP endpoints.

Quick setup

The README's quickstart needs Node.js 24.11 or newer. Then one command:

npx paperclipai@latest onboard --yes

The CLI runs from npm's cache, and your instance settings and data stay on your machine. For a local run, the README's FAQ says a single Node.js process manages an embedded PostgreSQL database and local file storage, so there's no separate database to set up. For production you can point it at your own Postgres.

If you'd rather build from source, the README's development steps are pnpm install then pnpm dev, which starts the UI and API at http://localhost:3100 (it needs pnpm 9.15+).

The full installation guide covers pinned versions, updates, rollback and uninstalling.

A simple first project for a small team

Keep the first run small and boring:

One goal, one or two agents. For example, "Clean up our website's FAQ page and open a pull request." Don't start with a company-wide goal.
Turn on review gates. Paperclip has configurable approval stages, so a human checks a plan before work starts.
Set a budget. The README says you can set company, agent and project budgets, get threshold alerts and pause work at a limit. Paperclip itself is free, but the AI models your agents call are not, so set the cap before you hit "go".
Read the task threads. Plans, blockers, files and run history stay attached to each task, which makes it easier to see what an agent actually did.

Things to check before going autonomous

Scope the secrets. The README mentions scoped secrets and permissions. Give each agent only the keys and repos it needs.
Start in a test repo. Let agents work on a copy before they touch anything customers see.
Keep a human on merges. "Approve and run" is the README's own workflow. Keep the approve part.

Is it worth trying?

If you're juggling several agent terminals and losing track of who's doing what, Paperclip is a free way to put that work on one board with budgets and approvals. If you only use one assistant now and then, it's probably more structure than you need today.

This is a research-based overview from the project's own README and docs. We haven't tested Paperclip hands-on yet.

More plain-English AI tool guides for small businesses and creators: ainews.useaifix.com. If you'd like help setting up AI automation for your business, see useaifix.com.

stageshow skill: How to Make Clean Product UI Demos With Your AI Coding Agent

AIFix — Sun, 01 Nov 2026 04:25:28 +0000

The stageshow skill is a new add-on for AI coding agents that helps them build "faux UI": simplified product screens, feature cards, bento grids and short cursor animations for landing pages and docs. It was released on 31 October 2026 by designer Benji Taylor, and if you run a small product or make content about software, it's worth a look. This is a research-based walkthrough based on the project's GitHub repo and the author's launch write-up.

What problem it solves

Most small teams need product visuals: a hero animation, a feature card, a "here's how it works" clip. You can ask an AI agent to mock one up, but as the author puts it, the results "often look sloppy and heavy-handed." Stageshow writes down the many small design decisions so the agent can make them too.

The skill is built around one rule from its SKILL.md: "It reads as a real product, simplified to what the viewer needs, framed with intent, and every motion has a cause." It also lists five common failures to avoid: too much detail, a bare wireframe, a whole window "boxed" into a card, a flat layout where nothing stands out, and a bland look that ignores the brief.

How to install it

From your project, run:

npx skills add benjitaylor/skills --skill stageshow

The author says this works with common coding agents such as Claude Code, Codex and Gemini. Then ask for something small: one interaction, in one product, at one size. The write-up's own example prompt is:

A small looping demo for our homepage hero: someone assigns an issue to a teammate.

The agent reads the skill, starts from the closest example, builds the demo, then films it and fixes what it finds. What you get is an ordinary React component you can drop into your site.

What's inside

SKILL.md: the core idea, the failure list and a seven-step workflow (medium and place, look, story, start from an example, and so on).
kit/: a small set of React parts (stage, windows, phones) that supports React 18 and later, with no third-party animation runtime.
examples/: dozens of "gold" examples, including studies of real products like Figma, Linear, Slack and Raycast.
scripts/: film.mjs records a demo and makes a contact sheet so the agent can check its own work, plus checks for one-frame jumps and text that doesn't fit.

There's also a local playground. It needs Node 20.19 or later:

npm install
npm run dev

A practical setup for a small team

Install the skill in the repo where your marketing site lives.
Pick one feature you explain often, like booking, invoicing or a dashboard view.
Ask for a 10 to 16 second loop (the length the skill recommends) at the size it will actually appear.
Ask for a compact phone version too; the skill designs a separate 358px-wide layout rather than squeezing the desktop one.
Review the motion yourself. The README is clear: "No automated clean result replaces watching the motion."

Honest caveats

The author warns it "still may not nail it in one shot," and most demos take a few rounds of feedback. These are illustrations, not automation: a scripted click only draws the cursor pressing. And the repo notes that each skill carries its own licence and third-party asset notices, so check those before using real brand marks.

We cover new AI tools for small businesses and creators in plain English at AI News. If you'd rather have the automation built for you, see AIFix.

antigravity vs claude code: A Practical Setup Guide for Small Dev Teams

AIFix — Sat, 31 Oct 2026 01:26:36 +0000

If you are a small team deciding on antigravity vs claude code, the most useful question is not "which is smarter?" but "which one fits how we already work?". Google Antigravity is an agent-first development platform with an app and a CLI. Claude Code is Anthropic's coding agent, which runs in your terminal, IDE, desktop app or browser. This research-based guide sticks to the official docs and covers setup, plans, plugins and team sharing.

Getting started

Claude Code. The overview docs list three install routes:

# macOS, Linux, WSL
curl -fsSL https://claude.ai/install.sh | bash

# Homebrew
brew install --cask claude-code

# Windows
winget install Anthropic.ClaudeCode

# then, inside your repo
cd your-project && claude

It also runs as an extension for VS Code and Cursor, as a JetBrains plugin, in a desktop app and on the web at claude.ai/code. Most surfaces need a Claude subscription or a Claude Console account.

Antigravity. Google's pricing page lists $0 a month for individuals, with business access through Google Cloud. The plans page says every plan includes Gemini 3.8 Flash, unlimited Tab completions, scheduled tasks and the CLI (agy).

Plans and limits

On Antigravity, Pro quota refreshes every 5 hours up to a weekly limit, while non-paid users get a weekly refresh. Pro and Ultra users can buy AI credits for overages, and there is no bring-your-own-key option. The models page lists Claude Sonnet 5.5 and Opus 5.5 on Pro (not during a trial), Ultra and Enterprise, alongside Gemini models. Older models, including Sonnet 4.6 and Opus 4.6, are removed on 2 November 2026.

On the Claude side, the pricing page shows Claude Code included from Pro ($20 a month, or $17 billed yearly) upward, but not on the Free plan.

Plugins: the setup that saves the most time

Both tools now bundle skills, MCP servers and hooks into plugins.

Claude Code (plugin docs):

# inside a Claude Code session
/plugin install commit-commands@claude-plugins-official

# from your shell
claude plugin list

Plugins can be installed at user, project (.claude/settings.json) or local scope. The docs warn that plugins can run hooks and MCP servers on your machine, so review them before installing.

Antigravity (plugin blog post, 29 October 2026): a plugin is a plugin.json manifest plus up to five parts: skills, MCP servers, rules, hooks and custom agents.

agy plugin install <plugin-name>
agy plugin validate ./<plugin-name>
agy plugin list --json

Already using Claude Code plugins? The Antigravity plugin docs include an importer that reads from ~/.claude/plugins/:

agy plugin import claude

Sharing setup across a team

Claude Code: commit a CLAUDE.md with your project rules (it can also read AGENTS.md) and keep shared plugins at project scope, so everyone gets the same setup on clone. The docs also cover GitHub Actions, GitLab CI and cloud scheduled tasks via /schedule.
Antigravity: workspace plugins live in .agents/plugins/ in the repo, and global ones in ~/.gemini/config/plugins/. Rules use rules/AGENTS.md inside a plugin.

A simple habit for both: keep one AGENTS.md with your build, test and style rules in the repo root, so switching tools later costs less.

A note for Jules users

Google's Jules changelog says Jules winds down on 31 December 2026, turns read-only after 11 November, and names Antigravity as Google's agent-first platform. Export your Jules data before then if you rely on it.

Which one to pick

Already paying for Claude, or living in the terminal: start with Claude Code.
Want a $0 starting point, or a Google Cloud setup: try Antigravity.
Not sure: keep your rules in AGENTS.md, run one real task in each, and compare.

We keep a research-based, side-by-side breakdown of plans, models and plugin support on ainews.useaifix.com. Small teams that want help setting up AI coding workflows can find us at useaifix.com.

claude code extension: How to Check if the Tampered Cursor Build Hit Your Machine

AIFix — Fri, 30 Oct 2026 13:26:09 +0000

If you use the claude code extension in Cursor, check your machine today. Cursor says a maliciously modified build of the extension, version 2.1.313, was available in its VS Code extension marketplace on 29 October 2026, and that at most 5,000 users were affected. Cursor also says this build "was not developed by Anthropic." This research-based guide sticks to what Cursor and Reuters have confirmed, and shows the quick check plus a few habits for locking down editor extensions.

What Cursor and Reuters confirmed

According to Cursor's security notice, the tampered build was live between about 2:27am and 1:35pm PT on 29 October. It:

collected host and user details such as hostname, username, OS, Git email and public IP
tried to add an attacker's SSH key to authorized_keys, turn on the local SSH service and open a reverse SSH tunnel
tried to make that tunnel survive restarts, using a LaunchAgent on macOS or a scheduled task on Windows

Cursor says turning on SSH needs admin rights, which the extension does not have on most macOS and Windows machines, so on those machines the attacker could not have logged in through the tunnel. The key, the persistence and the data upload can still happen, so Cursor asks affected users to clean up anyway. It also says the attacker's server is now offline, it found no evidence that other versions or the official distribution were affected, and it audited the marketplace and found no other tampered extensions.

Reuters, via CNA, reported the same points: up to 5,000 users, the same time window, and that Cursor shut down, rebuilt and relaunched its marketplace. Anthropic did not immediately respond to Reuters.

You may see louder claims online about who was behind it. We have left those out until they are confirmed.

Step 1: Check for the attacker's key

Cursor's notice says the build adds a public key labelled svc-telemetry. Its own check for macOS:

grep -qs svc-telemetry ~/.ssh/authorized_keys && echo "svc-telemetry key found" || echo "No key found; machine not impacted"

For Windows, Cursor gives a PowerShell check in the same notice that also looks at C:\ProgramData\ssh\administrators_authorized_keys. Copy it from there rather than retyping it.

Step 2: If the key is there

Follow Cursor's steps, in order:

Isolate the machine.
Run Cursor's remediation script for your platform (macOS as your normal user, Windows from an admin PowerShell). Cursor warns that uninstalling the extension may not be enough, because the payload leaves scripts that keep running.
Decide whether to rebuild the machine.
From a different, trusted computer, rotate passwords, API keys and sessions that the affected machine could reach.

Step 3: Check which version you have

In a VS Code-style editor, open the Extensions view and search @id:anthropic.claude-code to find the exact extension and its version. VS Code's extension docs also list code --list-extensions --show-versions for the terminal. Cursor names the bad folders anthropic.claude-code-2.1.313-darwin-arm64 and anthropic.claude-code-2.1.313-win32-x64.

Step 4: Lock down your editor extensions

These settings come from VS Code's documentation. Check that your editor version supports them before you rely on them.

Slow down updates. extensions.autoUpdateDelay waits a set number of hours (12 by default) before an update installs automatically. You can also turn auto-update off for the extensions you care most about.
Allow only what you need. extensions.allowed blocks any extension not on your list. That works well for a small team.
Prune. Uninstall extensions you no longer use. Each one runs with the same permissions as the editor.
Install by exact ID. Search with @id:publisher.extension to avoid look-alike names.

The bottom line

The claude code extension incident was short, and Cursor says most machines could not be reached through the tunnel. The one-line check takes seconds, so run it, and treat editor extensions like any other software with access to your files.

We share plain, research-based AI news for small businesses and creators at ainews.useaifix.com, and AI automation help at useaifix.com.