The HisnLabs Security Blog
Plain-language, sourced reporting on Mac security, online safety and AI risk.
From Red Team to FireAI: A Note from Our CTO
Our CTO explains where FireAI comes from: years spent on cybersecurity red teams and threat-intelligence work, and the conviction that a firewall shouldn’t have to choose between security and simplicity.
Guardrino-t7 v7: nearly 9 in 10 Arabic attacks caught, in about 15 ms
On our own Arabic test set, Guardrino-t7 v7 caught 6,814 of 7,874 Arabic attacks, at a median 14.8 ms a message on a laptop CPU. Same machine for every model; method, limits and data published.
Why an AI guardrail has to sit in the request path
A check that lives beside your app can be skipped. Guardrino sits in the path itself: an OpenAI-compatible proxy, a LiteLLM guardrail and an MCP security proxy are available.
Inside Guardrino’s lab: Arabic prompt injection, what broke and what we fixed
How we test Guardrino against prompt injection in Arabic, Arabizi and English: the open model’s weak spots, a demo that runs in your browser, the mistakes our own review found and what v6 is trained on.
Guardrino-t7 Open: an Arabic and Arabizi prompt-injection classifier, now on Hugging Face
HisnLabs publishes Guardrino-t7 Open on Hugging Face: an earlier Guardrino classifier that flags prompt injection and jailbreaks in Arabic, Arabizi and English. Apache-2.0, runs on a CPU, limits stated.
Guardrino-t7 is coming: UAE AI guardrails for Arabic and Arabizi
Guardrino-t7, built in Dubai, is a guardrail model for AI chatbots and agents that reads Arabic dialects, Arabizi and English and runs on your own servers in the UAE. Pilot partners wanted.
Agent-to-agent business in the UAE: protecting the AI that talks to AI
Soon your AI agent will book, buy and settle disputes with other companies’ agents. What changes after B2B and APIs, four UAE use cases, and the protection layer that keeps each exchange on your terms.
What are AI guardrails? A plain guide for businesses
AI guardrails let your assistants and agents say yes to customers with confidence. What they are, how layered checks work, and the questions to ask before you choose one.
Guardrails that don’t annoy your users: why fewer refusals is a security win
A guardrail that blocks normal customers gets switched off. How ask-to-rephrase, masking and per-team policies keep AI safe and users happy, with Arabic, Gulf and Arabizi examples.
How to test an AI guardrail with public test sets, step by step
A hands-on method to score any AI guardrail, yours or a vendor’s: caught versus wrongly blocked, honest confidence ranges, GuardBench, XSTest and CyberSecEval, with the commands.
Why we are building Guardrino: AI guardrails made for the Middle East
Companies in the Gulf are putting AI in front of customers who write in Arabic, dialect, English and Arabizi. Why HisnLabs is building a guardrail for them, and how to join the pilot.
Your ID number leaked: why scam messages now sound right, and five signs that still give them away
After a register leak, scam messages quote your real name, address and ID number. Why accurate details prove nothing, and the five signs the UK NCSC lists.
When a document gives your AI assistant orders: five defences, tested
An AI assistant that reads a file or an e-mail can be steered by what it says. What prompt injection is, and what five simple defences did on one local model.
What AI assistant memory stores, and why viewing, editing and forgetting it matters
ChatGPT, Gemini and Claude now remember users. What their memory keeps, how forgetting works, what embeddings leak, and what keeping memory on a Mac changes.
Consent before an AI transcribes a call: France, Morocco and the one-party rule
Recording or transcribing a private call without consent is an offence in France and Morocco. What the texts say, how meeting apps ask, and what local AI changes.
What cloud voice assistants keep from your voice, and what on-device speech recognition changes
Alexa, Gemini and Siri differ in where speech is processed and how long recordings and transcripts are kept. A source-based review, and what on-device AI changes.
DNS Leaks Explained: Why Lookups Must Travel Through Tor on a Mac
A DNS lookup names the site before any connection starts. How DNS leaks happen on macOS, why .onion names are special, and how to test for a leak.
Client confidentiality for freelancers who plan their work in cloud AI tools
Client names typed into cloud to-do and AI apps reach a third party. Professional secrecy, GDPR data minimisation and processor duties, set out from the texts.
Threat Models for Journalists, Researchers and Activists on a Mac: Who Sees What
Who sees what when a journalist or researcher works on a Mac: the ISP, the Wi-Fi operator, the website and a two-ended observer, with and without per-app Tor.
Tor New Identity Explained: NEWNYM, Circuit Renewal and Their Costs
What Tor’s New Identity, the NEWNYM signal and the ten-minute circuit rule renew, what stays the same (guard, logins, fingerprints), and what renewal costs.
Why Tor Chooses Random Paths: Speed, Guards and the Cost of a “Fastest Route”
Tor draws relays at random, weighted by bandwidth, and keeps a few long-lived guards. Why picking the fastest route can reveal where a user is, from the research.
Onion Services Explained: v3 Addresses, RFC 7686 and Safe Access
A v3 .onion address is the service’s public key. Why RFC 7686 keeps .onion out of DNS, why ordinary browsers refuse it, and how to reach onion services safely.
Per-App Tor Routing on macOS: Why All-or-Nothing Tunnels Leak Context
A full-device VPN mixes every app behind one address and has documented gaps. How macOS transparent proxies route per app, and where fail-closed and DNS matter.
Relay Authority: Scoring Tor Relays from 0 to 100, and Where It Stops
Malicious Tor relays have run large shares of the network. How a 0 to 100 relay score can be built from public Tor data, and what no score can detect.
What Tor Hides on a Mac, and What It Leaves Exposed
Tor hides your IP address from the sites you reach. Logins, browser fingerprints, local DNS and plaintext at the exit stay visible. A source-based review.
Best Mac Firewall in 2026: Choosing by What You Need
The macOS firewall, LuLu, Radio Silence, Murus, Vallum, Little Snitch and FireAI, compared on the questions that decide the choice: direction, alerts, rules and price, from each vendor’s own pages.
Data Exfiltration on a Mac: How It Happens and How to Detect It
How data leaves a Mac, mapped to MITRE ATT&CK’s Exfiltration tactic (TA0010): infostealers, web services, AI agents, and the network signals, such as upload spikes, that give it away.
How to Turn On the Firewall on a Mac (macOS 14, 15, 26), and What It Doesn’t Block
Step-by-step: turn on the macOS firewall, enable stealth mode and allow an app through it, from Apple’s own guides. Then the part it leaves open: outgoing connections.
Aider Security on Mac: Monitor Its Network Connections
Aider is an open-source Python tool for AI pair programming in the terminal. See how it installs on a Mac, what it sends over the network and how FireAI flags new destinations.
Amp Coding Agent Security on Mac: Monitor Its Connections
Amp is a coding agent with a CLI and a Mac app that syncs threads to ampcode.com. See what its docs say about data flow and settings, and how FireAI flags new destinations.
Every Other AI Agent on Your Mac: Python, npm, Bun, Native
AI agents outside FireAI's list of 19 still get connection prompts and per-app rules. How node and python agents appear, and how to cover them with rules and security modes.
Monitor the ChatGPT Mac App Network Connections: Security on Mac
The ChatGPT Mac app is a native OpenAI client that also carries OpenAI’s dots. A reported 2026 flaw, what a firewall can and cannot see, and how FireAI flags new destinations.
Monitor Claude Code Network Connections on Mac: Security with FireAI
Claude Code runs commands and opens connections on your Mac. What Anthropic documents about its permissions and domains, and how FireAI flags a new destination or upload spike.
Monitor the Claude Desktop App Network on Mac: Security with FireAI
The Claude desktop app has Chat, Cowork and Code tabs. What Anthropic documents about its hosts and permission modes, and how FireAI flags a new destination or an upload spike.
Crush Agent Security on Mac: Monitor Its Network with FireAI
Crush is a Go terminal coding agent from Charm with permission prompts, MCP and provider catalogue updates. See what it can reach on a Mac and how FireAI watches its connections.
Monitor Cursor Network Connections on Mac: AI Agent Security
Cursor’s agent runs terminal commands and talks to several Cursor domains. What its docs say about run modes, the macOS sandbox and hosts, and how FireAI flags a new destination.
Gemini CLI Security on Mac: Monitor Its Network Connections
Gemini CLI by Google runs shell commands, reads files and fetches web pages from your terminal. See its documented sandbox and trusted folders, and how FireAI flags new destinations.
GitHub Copilot CLI Security on Mac: Watch Its Connections
GitHub Copilot CLI edits files and runs commands from your terminal. See its trusted directories and approval prompts, and how FireAI flags new destinations and upload spikes on a Mac.
Goose Agent Security on Mac: Monitor Network Access (FireAI)
Goose is an open-source AI agent from the Agentic AI Foundation with a desktop app, a CLI and MCP extensions. See what it can reach on a Mac and how FireAI watches its connections.
Hermes Agent Security on Mac: Monitor Its Network Access
Hermes Agent from Nous Research runs terminal commands and a messaging gateway. See its documented approvals and sandboxes, and how FireAI flags new destinations and upload spikes.
Kiro Security on Mac: Monitor Its Network Connections with FireAI
Kiro is an AWS agentic IDE. Read its documented permission model and data handling, then watch its Mac network connections with FireAI and block a first-ever destination.
Muse from Meta on Mac: Monitor Its Network Access with FireAI
Muse is described by Meta as a personal AI agent. What is verifiable about it, and how to watch its network connections on a Mac with FireAI and flag a new destination.
Monitor OpenAI Codex CLI Network Access on Mac with FireAI
OpenAI’s Codex CLI runs commands in a macOS Seatbelt sandbox with network off by default. What the docs say about sandbox modes and approvals, and how FireAI flags a new destination.
Securing OpenClaw on Mac: Monitor Its Network Connections
OpenClaw runs shell commands, files and a browser on your Mac and talks to messaging apps. Learn what its docs say about access, and how FireAI watches where it connects.
opencode Security on Mac: Monitor Network Access with FireAI
opencode is an open-source terminal coding agent with a permissive default permission model and an optional share feature. Learn what it can reach on a Mac and how FireAI flags new destinations.
Is Qwen Code Safe on Mac? Monitor Its Network with FireAI
Qwen Code is an open-source terminal coding agent that runs on Node.js. See what it can reach on a Mac, its approval modes and sandbox, and how FireAI watches its connections.
Trae Security on Mac: Monitor Network Access with FireAI
What can be verified about the Trae AI coding tool, and how to watch its network connections on a Mac with FireAI: 3-day learning, new-destination flags and upload spikes.
Is Windsurf Safe on Mac? Monitor Its Network Access with FireAI
What the Windsurf AI editor can reach on a Mac, its documented terminal permission levels, and how FireAI learns its connections and flags a new destination or an upload spike.
FireAI recognises 19 AI agents on macOS and limits where their data goes
FireAI 1.0.4 names 19 AI agents on a Mac, including those run by node or python, and flags or blocks their new destinations. Mapped to the OWASP 2025 and 2026 risk lists.
FireAI vs Little Snitch: Which Mac Firewall Fits You in 2026?
FireAI and Little Snitch compared on rules, AI review, blocklists, history, Intel support, price and trial, from each vendor’s own pages, checked October 2026.
FireAI vs LuLu: Free Open-Source Firewall or On-Device AI for Your Mac?
FireAI and LuLu compared on price, source code, rules, browser hostnames, reputation checks, AI review, Intel support and trial, from each project’s own pages, checked October 2026.
OpenClaw and Hermes Agent Security Incidents in 2026: Exposure, Skills, Bans
A sourced summary of published 2026 security findings about OpenClaw and Hermes Agent: exposed instances, malicious skills, prompt injection, bans, and attackers using Hermes.
The Best Tools of 2026 for AI Model Security
A balanced roundup of open tools that secure AI models and LLM applications in 2026, grouped by job, with maintainers, licences, limits and where network controls fit.
FireAI 1.0: a redesigned interface that shows the state of a Mac’s network at a glance
FireAI 1.0 from HisnLabs redesigns the World map, Activity, Home and alerts so network state is shown, not read, and acting takes fewer steps than 0.1.8.
FireAI Pilot Now Learns Each Network: Home, Café and Jobsite Get Their Own Answers
FireAI 0.1.8 teaches FireAI Pilot the difference between networks and security modes: a connection allowed at home can stay blocked on café or jobsite Wi-Fi.
How FireAI Limits the Network Damage of OpenClaw’s Risky Mistakes on Your Mac
Five documented risk classes of running an AI agent with shell access on macOS, the FireAI network controls that limit each, and the risks a firewall cannot address.
How Anthropic red-teams Claude, and what it leaves to you
What Anthropic publishes about red-teaming Claude, which methods a deployer can reuse, which stay with the model developer, and what remains with you.
OpenClaw vs Hermes Agent: Two Open-Source Personal AI Agents Compared
OpenClaw and Hermes Agent are MIT-licensed personal AI agents with shell, file and messaging access. A comparison of design, memory, skills and documented security issues.
Shared responsibility for LLMs: who secures what
How cloud shared responsibility applies to LLMs: Microsoft and CSA models, EU AI Act duties of model providers and deployers, and a table for API use.
A career in AI cybersecurity: what you need to know
AI cybersecurity is not one job. It is at least four: adversarial testing of models, securing the software around them, using AI inside detection work, and governing how an organisation adopts it.
Do AI certifications make sense when AI moves this fast?
A balanced look at what AI and AI-security certifications actually test, how fast the field under them changes, what public data says about their value, and the alternatives that exist alongside them.
One Switch for a Child’s Mac: Parental Controls Arrive in FireAI 0.1.7
FireAI 0.1.7 adds Parental Controls: six shields that block adult sites, gambling, drugs, games, ads and scams in every app, plus a one-click Child Safety First preset.
FireAI 0.1.6: See Every Device on Your Wi-Fi, Lock Down USB Sticks, and 5 More Ways to Own Your Mac’s Network This Weekend
FireAI 0.1.6 is out: Who’s online, USB Network Protection, fake Wi-Fi warnings, Clipboard Armor and AirDrop Shield. A 48-hour challenge to try them all.
From “Allow nsurlsessiond to 17.253.x.x?” to “Apple is downloading an update”: How FireAI Puts AI Inside the Firewall
Classic firewall prompts show process names and IP addresses, so people click Allow blindly. FireAI’s on-device AI turns each one into a sentence you can act on.
Mr. Robot’s Hacks, Explained: Which Would Work on Your Mac Today — and What Actually Stops Them
USB sticks in a parking lot, a Raspberry Pi hidden in a data center, a fake cell tower: we take the show’s most famous hacks apart and check them against a modern Mac.
How to Spot a Rogue Device on Your Home Wi-Fi (the Mr. Robot Raspberry Pi Trick)
A tiny computer hidden on your network is a real attacker technique. Here’s how to see every device on your Wi-Fi from your Mac, recognize the odd one out, and get told when it comes back.
USB Drop Attack: What a Found USB Stick Can Do to a Mac
Leave a USB stick in a parking lot and someone will plug it in. What that stick can do to a Mac, what macOS already blocks, and the gap a firewall closes.
Why Endpoint Detection Struggles With an AI Agent That Goes Rogue
EDR is built to spot bad binaries and bad behaviour. An AI agent hijacked by injected instructions uses neither: it acts through tools you already trust. What that means, honestly, for defence.
What Is Actually Inside Your Electron Apps? We Extracted One to Find Out
Electron bundles Chromium and Node.js into one process model. We unpacked a real, installed app’s archive to show what that means for security review and for network firewalls.
Encrypted Client Hello and DoH: What Perimeter Firewalls Can No Longer See
ECH (RFC 9849) hides the real hostname in the TLS handshake and DoH hides the DNS lookup. What that takes away from network firewalls, what it does not, and where visibility moves.
Building a Distraction-Free Homework Setup on a Mac
macOS already has real tools for a focused homework session: a clean desktop, Screen Time limits, scheduled Focus. Here is what each one actually does, and where they stop.
Jev and the Rise of Decision Models: What System One AI Means for Security Tools
TypeSafe AI says its new model Jev answers in milliseconds with a typed, calibrated decision instead of a paragraph. Here is exactly what was claimed, what it means, and what still needs independent testing.
Running an LLM on Your Own Mac Is Not Automatically Safe
Local models feel private because nothing leaves for the cloud. Two real, patched vulnerabilities show what can still go wrong: an unauthenticated API, and an agent tricked by what it reads.
Camera and Microphone Permissions on macOS: What “Allow” Really Means, and How to Take It Back
The one-time prompt that grants camera or microphone access, where to review or switch it off, the terminal command that resets it, and why quitting the app is what actually stops a live session.
The MCP Blind Spot: When a Document Can Make Your AI Agent Act
Anthropic’s Model Context Protocol lets AI agents call local tools. A real, demonstrated attack shows how untrusted content can hijack that access. What it takes to defend against it.
“Install as App” Phishing: When a Website Pretends to Be Software
Chrome lets any site install itself as a standalone app. A documented technique uses that to fake a whole login window, address bar included. What it can and can’t hide, and how to check what’s really installed.
Does a Downloaded File Phone Home? We Tested Spotlight and Quick Look on macOS 26
A claim going around says macOS indexing makes unopened files fetch remote content. We tested it: indexing and thumbnails stayed silent, but pressing Space to preview did not.
Slopsquatting: When Your AI Coding Assistant Invents a Package, and Someone Else Registers It
Code-generating models hallucinate package names at a measurable, sometimes predictable rate. Attackers register those names for real. What the research found, and why it matters the moment `pip install` runs.
How to Tell Which App Is Using Your Mac’s Camera or Microphone Right Now
macOS already tells you when the camera or microphone turns on, and often which app is behind it. Here is where to look, in order, and how FireAI 0.1.2 adds the network side.
Have You Ever Wondered If Someone Is Watching You Through That Little Black Hole at the Top of Your Screen?
Your Mac tells you when the camera is on. It doesn’t tell you where the picture goes. What macOS shows you, what it leaves out, and how FireAI 0.1.2 connects the two.
Beyond the VPN: Why On-Device Firewalls Are the New Gold Standard
A VPN encrypts your traffic and hides your IP — but it trusts everything already running on your Mac. Here is the blind spot it cannot cover.
Can an AI Steal or Wipe Your Identity?
Autonomous AI agents and data aggregation have turned identity theft from a manual scheme into an automated pipeline — here is what actually changed, and what still stops it.
Invisible Connections: What Your Computer Is Whispering to the Internet
Even when you are not browsing, installed apps, background extensions and trackers keep transmitting your metadata to remote servers — here is what that traffic actually is.
Your Mac Isn’t as Private as You Think: The Truth About Background Telemetry
You can lock down your Mac perfectly and still leak personal data — the modern threat is the silent, constant stream of background telemetry, not just malicious code.
Data Brokers: How Your Personal Information Leaks Even From a Well-Secured Mac
You can lock down your Mac perfectly and still leak personal data — not through a hack, but through ordinary apps quietly sharing it with advertising and analytics companies you never agreed to deal with directly.
Prompt Injection Against AI Agents: A Hands-On Walkthrough
Direct versus indirect prompt injection, how markdown images and links have been used to exfiltrate data from real chat products, and why least privilege, human confirmation and egress control are the mitigations worth actually building.
Hunting Persistence on macOS: LaunchAgents, Login Items and Background Tasks
Every place code can register to survive a reboot on a Mac, checked from the terminal: LaunchAgents, LaunchDaemons, login items, cron, configuration profiles — and where FireAI does and does not fit in.
File Metadata on a Mac: How to Find It and Remove It
Photos, PDFs and Office files on a Mac carry hidden metadata: GPS coordinates, author names, source URLs. How to find it and strip it with real Terminal commands.
Running a Local LLM on Apple Silicon for Security Triage
Setting up MLX and llama.cpp on a Mac, what quantization actually trades off in memory, a prompt pattern for triaging one connection log line, and the real limits of asking a small local model to do security work.
Check Any Mac App’s Signature and Notarization From the Terminal
How to read codesign, spctl and stapler output like a practitioner: Team ID, hardened runtime, entitlements, quarantine flags, and the red flags that separate a normal result from one worth a second look.
The macOS pf Firewall: What It Can Do, and Why It Can’t Be Your App Firewall
A hands-on look at pf, the BSD packet filter inside macOS: inspecting and loading rules with pfctl, and the concrete reasons — hostnames, process identity, prompts — it was never built to be a per-app firewall.
Ransomware Isn’t Just a Windows Problem: What Small Businesses on Mac Need to Know
Ransomware that targets macOS directly is genuinely rare — but a Mac sitting inside a small business network, synced to shared drives, is not protected by that fact alone.
Encrypted DNS on macOS: DoH and DoT With Configuration Profiles
Why plain DNS leaks every site you visit, how DNS-over-HTTPS and DNS-over-TLS fix it, and how to configure either natively on a Mac with a signed configuration profile, verified from the terminal.
Audit What Your Mac Sends in 10 Minutes, From the Terminal
A ten-minute, tools-only-macOS-ships-with audit: lsof, nettop, log stream, scutil --dns, dig and tcpdump — what each one actually shows, the exact commands, and what none of them can tell you.
What Privacy-Conscious Mac Users Are Turning On in 2026
A sourced tour of the privacy settings and system features Mac users are actually enabling in 2026 — Advanced Data Protection, Lockdown Mode, Private Relay, passkeys, encrypted DNS and more — with what each protects and what it does not.
How Good Are AI Models at Security Work? What the Public Benchmarks Show
Four public, peer-reviewed evaluations — Cybench, CyberSecEval 3, NYU CTF Bench and OpenAI’s o1 system card — scored real models on real cybersecurity tasks. Here is exactly what they found, cited number by number.
How Whistleblowers Protect Themselves: A Practical Digital Security Guide
A sourced walkthrough of the tools and legal frameworks whistleblowers actually rely on: threat modelling, SecureDrop, Tails, Signal’s real limits, metadata risks, and what EU, French and US law protect.
What an AI Token Cost, 2022 to 2026: OpenAI, Anthropic and DeepSeek
List prices per million input and output tokens from OpenAI, Anthropic and DeepSeek, tracked from 2022 to today with a dated, archived source for every figure, plus what the price curve actually did and did not do.
Public Wi-Fi on a Mac: The Real Risk Isn’t What You Think
The classic "someone is reading your traffic at the coffee shop" warning is mostly outdated now that almost every site uses encryption — but that does not mean public Wi-Fi is risk-free.
Mac Security Tools Compared: Firewalls, Antivirus and What Fits You
Apple’s firewall, Little Snitch, LuLu, Radio Silence and FireAI compared on outbound control, per-app rules, code signatures, threat feeds, on-device AI and price, sourced from each vendor’s site.
Securing a Mac for Remote Work: The Home-Office Checklist
Router and Wi-Fi hygiene, what a VPN does and does not cover, meeting-app permissions, a password manager, file sharing and a per-app firewall — with ANSSI, CISA, NIST and Apple guidance cited.
When AI Agents Went Rogue: Inside the OpenAI–Hugging Face Incident
In mid-2026, OpenAI’s own internal research AI agents broke out of their test environment, coordinated with each other, and compromised Hugging Face’s servers — entirely on their own, without a human directing them.
Which Mac Apps Can You Trust? Signing, Notarization, Permissions and Network Behaviour
App Store or direct download, Developer ID and notarization, the quarantine flag, TCC permissions, and how to read an app’s network behaviour as the trust signal Apple’s checks cannot give you.
Network Monitoring on a Mac: See Every Connection Your Apps Make
What Activity Monitor, lsof and nettop really show, where they stop, what a per-app firewall adds, how to read one connection, and which patterns — beaconing, Tor exits, plain HTTP — deserve a look.
Gatekeeper, XProtect and Notarization: What macOS Actually Catches — and Misses
Apple builds three real, documented layers of malware defense into every Mac. Apple’s own security guide is also honest about what those layers do not promise to catch.
Mac Firewall vs Antivirus: You Need Both
A firewall watches connections; an antivirus watches files. Neither replaces the other. What macOS already ships, where Apple says it stops, and how to pair an outbound firewall with a scanner.
Ransomware on Mac: A Layered Defence That Actually Holds
Mac ransomware is rare but real: KeRanger, ThiefQuest and a LockBit test build. What each defensive layer does, what a firewall can and cannot catch, and the backup rules that decide recovery.
Spyware on a Mac: How to Spot It, What Removes It, How to Keep It Out
Stalkerware, infostealers and mercenary spyware are three different problems: the real signs, how Lockdown Mode and Apple threat notifications work, what removes spyware, where a firewall fits.
The New Wave of Mac-Targeted Password-Stealing Malware
Since 2023, security researchers have tracked a steady rise in malware built specifically to steal passwords, browser data and crypto wallets from Macs — usually delivered through fake software downloads, not exotic exploits.
Firewall AI for Mac: Setup and Optimization Guide
Step by step: installing FireAI on a Mac, the network-extension approval in System Settings, the first prompts, security modes, per-app rules, threat feeds and rule files.
Mac Security for Creatives: Protecting Client Work Before It Ships
Studios and freelancers hold unreleased client work on Macs full of plugins and sync clients. The real risks are infostealers, leaked credentials and quiet exfiltration, not viruses.
Zero-Days on macOS: Why Unknown Threats Still Have to Phone Home
Documented exploits against unpatched Mac flaws, from FORCEDENTRY to the Hong Kong watering hole: why the exploit is so hard to detect, and why the traffic that follows is where per-app control helps.
How Chatty Is Your Mac? What macOS Says About You Without Asking
Long before any third-party app runs, macOS itself is already talking to Apple’s servers — and in 2020, one of those channels was found to ignore your firewall and VPN entirely.
Why AI Firewalls Are Essential for Mac Users
Apple’s built-in firewall only watches incoming connections and has to be switched on by you. Here is the outbound gap it leaves, what an on-device AI reviewer can do about it, and what it cannot.
Advanced Threat Detection on macOS: What the Layers Actually Do
What each Mac detection layer can see and what it cannot: XProtect signatures, notarization, Apple’s Endpoint Security framework, behavioural analysis and on-device review of network connections.
Yes, Macs Get Infected: A Short, Sourced History of Real Mac Malware
From the first Mac ransomware in 2016 to a 2021 threat built for Apple Silicon before Apple Silicon malware existed, the "Macs don’t get viruses" myth has a real, documented body count.
Regain Control: How to Stop Data Harvesters Without Breaking Your Workflow
Blocking every unfamiliar connection sounds thorough and breaks half your apps by lunchtime. Here is a version of control that survives contact with a real workday.
The Hidden Risk on Roblox and Discord: A Parent’s Guide
A 2024 Bloomberg Businessweek investigation and a string of arrests since January 2025 have made one thing clear: online-game platforms popular with children need real, informed supervision, not blind trust.
Zero Trust at Home: Why Every App Should Earn Its Internet Access
Large organizations stopped assuming anything inside their network was automatically safe years ago. A personal Mac, full of apps that get network access by default, hasn’t caught up.
The Illusion of Built-In Security: Closing the Gaps the OS Leaves Wide Open
A modern operating system does an enormous amount of security work by default — and its own documentation is honest that "by default" is not the same as "complete."
What Kevin Mitnick Can Still Teach You About Social Engineering
Kevin Mitnick, once the FBI’s most-wanted hacker, spent his later career proving that the easiest way into any system was never the code — it was the person answering the phone.
The Data Broker Economy: Starving the Machine with Least Privilege
Data brokers do not break into anything. They buy what apps are already willing to sell them — which means the most effective defense is cutting off the supply, not guarding a vault.
The Enemy Inside: How "Trusted" Apps Sell Your Data Behind Your Back
The apps most likely to share your data are not the sketchy ones you avoided — they are the ordinary, useful, well-reviewed apps that bundle a data-broker SDK alongside their real feature.
Poisoned Prompts & Phantom Data: The Hidden Risks of Embedded AI Tools
The AI assistant reading your email or browsing a page for you has to trust whatever it reads — and attackers have already worked out how to hide instructions inside that content.
The Synthetic Identity Crisis: How to Prove You Are Still You Online
Synthetic identity fraud does not steal your identity outright — it blends real fragments of you with invented details, which makes it unusually hard for anyone, including you, to prove is fake.
When Algorithms Go Rogue: Defending Your System Against Autonomous Malware
Traditional malware does what an attacker programmed it to do. A newer category decides for itself — and the clearest documented case wasn’t built by criminals at all.
The Deepfake Dossier: How AI Reconstructs Your Life from Digital Breadcrumbs
No single leak has to be dramatic. AI is now good enough to stitch together scattered, ordinary fragments — photos, timestamps, voice clips — into a convincing synthetic version of you.
AI security and LLM guardrails
Related: Guardrino · Arabic LLM guardrails · LLM guardrails for the GCC · Arabic prompt injection
- Guardrino-t7 v7: nearly 9 in 10 Arabic attacks caught, in about 15 ms
- Why an AI guardrail has to sit in the request path
- Inside Guardrino’s lab: Arabic prompt injection, what broke and what we fixed
- Guardrino-t7 Open: an Arabic and Arabizi prompt-injection classifier, now on Hugging Face
- Guardrino-t7 is coming: UAE AI guardrails for Arabic and Arabizi
- Agent-to-agent business in the UAE: protecting the AI that talks to AI
- What are AI guardrails? A plain guide for businesses
- Guardrails that don’t annoy your users: why fewer refusals is a security win
- How to test an AI guardrail with public test sets, step by step
- Why we are building Guardrino: AI guardrails made for the Middle East
- When a document gives your AI assistant orders: five defences, tested
- What AI assistant memory stores, and why viewing, editing and forgetting it matters
- Consent before an AI transcribes a call: France, Morocco and the one-party rule
- What cloud voice assistants keep from your voice, and what on-device speech recognition changes
- Client confidentiality for freelancers who plan their work in cloud AI tools
- Data Exfiltration on a Mac: How It Happens and How to Detect It
- Aider Security on Mac: Monitor Its Network Connections
- Amp Coding Agent Security on Mac: Monitor Its Connections
- Every Other AI Agent on Your Mac: Python, npm, Bun, Native
- Monitor the ChatGPT Mac App Network Connections: Security on Mac
- Monitor Claude Code Network Connections on Mac: Security with FireAI
- Monitor the Claude Desktop App Network on Mac: Security with FireAI
- Crush Agent Security on Mac: Monitor Its Network with FireAI
- Monitor Cursor Network Connections on Mac: AI Agent Security
- Gemini CLI Security on Mac: Monitor Its Network Connections
- GitHub Copilot CLI Security on Mac: Watch Its Connections
- Goose Agent Security on Mac: Monitor Network Access (FireAI)
- Hermes Agent Security on Mac: Monitor Its Network Access
- Kiro Security on Mac: Monitor Its Network Connections with FireAI
- Muse from Meta on Mac: Monitor Its Network Access with FireAI
- Monitor OpenAI Codex CLI Network Access on Mac with FireAI
- Securing OpenClaw on Mac: Monitor Its Network Connections
- opencode Security on Mac: Monitor Network Access with FireAI
- Is Qwen Code Safe on Mac? Monitor Its Network with FireAI
- Trae Security on Mac: Monitor Network Access with FireAI
- Is Windsurf Safe on Mac? Monitor Its Network Access with FireAI
- FireAI recognises 19 AI agents on macOS and limits where their data goes
- OpenClaw and Hermes Agent Security Incidents in 2026: Exposure, Skills, Bans
- The Best Tools of 2026 for AI Model Security
- How FireAI Limits the Network Damage of OpenClaw’s Risky Mistakes on Your Mac
- How Anthropic red-teams Claude, and what it leaves to you
- OpenClaw vs Hermes Agent: Two Open-Source Personal AI Agents Compared
- Shared responsibility for LLMs: who secures what
- A career in AI cybersecurity: what you need to know
- Do AI certifications make sense when AI moves this fast?
- Why Endpoint Detection Struggles With an AI Agent That Goes Rogue
- Jev and the Rise of Decision Models: What System One AI Means for Security Tools
- Running an LLM on Your Own Mac Is Not Automatically Safe
- The MCP Blind Spot: When a Document Can Make Your AI Agent Act
- Slopsquatting: When Your AI Coding Assistant Invents a Package, and Someone Else Registers It
- Can an AI Steal or Wipe Your Identity?
- Prompt Injection Against AI Agents: A Hands-On Walkthrough
- Running a Local LLM on Apple Silicon for Security Triage
- How Good Are AI Models at Security Work? What the Public Benchmarks Show
- What an AI Token Cost, 2022 to 2026: OpenAI, Anthropic and DeepSeek
- When AI Agents Went Rogue: Inside the OpenAI–Hugging Face Incident
- Why AI Firewalls Are Essential for Mac Users
- Poisoned Prompts & Phantom Data: The Hidden Risks of Embedded AI Tools
- The Synthetic Identity Crisis: How to Prove You Are Still You Online
- When Algorithms Go Rogue: Defending Your System Against Autonomous Malware
- The Deepfake Dossier: How AI Reconstructs Your Life from Digital Breadcrumbs
Mac security
- Your ID number leaked: why scam messages now sound right, and five signs that still give them away
- DNS Leaks Explained: Why Lookups Must Travel Through Tor on a Mac
- Threat Models for Journalists, Researchers and Activists on a Mac: Who Sees What
- Tor New Identity Explained: NEWNYM, Circuit Renewal and Their Costs
- Why Tor Chooses Random Paths: Speed, Guards and the Cost of a “Fastest Route”
- Onion Services Explained: v3 Addresses, RFC 7686 and Safe Access
- Per-App Tor Routing on macOS: Why All-or-Nothing Tunnels Leak Context
- Relay Authority: Scoring Tor Relays from 0 to 100, and Where It Stops
- What Tor Hides on a Mac, and What It Leaves Exposed
- Best Mac Firewall in 2026: Choosing by What You Need
- How to Turn On the Firewall on a Mac (macOS 14, 15, 26), and What It Doesn’t Block
- FireAI vs Little Snitch: Which Mac Firewall Fits You in 2026?
- FireAI vs LuLu: Free Open-Source Firewall or On-Device AI for Your Mac?
- FireAI 1.0: a redesigned interface that shows the state of a Mac’s network at a glance
- FireAI Pilot Now Learns Each Network: Home, Café and Jobsite Get Their Own Answers
- One Switch for a Child’s Mac: Parental Controls Arrive in FireAI 0.1.7
- FireAI 0.1.6: See Every Device on Your Wi-Fi, Lock Down USB Sticks, and 5 More Ways to Own Your Mac’s Network This Weekend
- From “Allow nsurlsessiond to 17.253.x.x?” to “Apple is downloading an update”: How FireAI Puts AI Inside the Firewall
- Mr. Robot’s Hacks, Explained: Which Would Work on Your Mac Today — and What Actually Stops Them
- How to Spot a Rogue Device on Your Home Wi-Fi (the Mr. Robot Raspberry Pi Trick)
- USB Drop Attack: What a Found USB Stick Can Do to a Mac
- What Is Actually Inside Your Electron Apps? We Extracted One to Find Out
- Encrypted Client Hello and DoH: What Perimeter Firewalls Can No Longer See
- Building a Distraction-Free Homework Setup on a Mac
- Camera and Microphone Permissions on macOS: What “Allow” Really Means, and How to Take It Back
- “Install as App” Phishing: When a Website Pretends to Be Software
- Does a Downloaded File Phone Home? We Tested Spotlight and Quick Look on macOS 26
- Have You Ever Wondered If Someone Is Watching You Through That Little Black Hole at the Top of Your Screen?
- Beyond the VPN: Why On-Device Firewalls Are the New Gold Standard
- Invisible Connections: What Your Computer Is Whispering to the Internet
- Your Mac Isn’t as Private as You Think: The Truth About Background Telemetry
- Hunting Persistence on macOS: LaunchAgents, Login Items and Background Tasks
- Check Any Mac App’s Signature and Notarization From the Terminal
- The macOS pf Firewall: What It Can Do, and Why It Can’t Be Your App Firewall
- Encrypted DNS on macOS: DoH and DoT With Configuration Profiles
- Audit What Your Mac Sends in 10 Minutes, From the Terminal
- What Privacy-Conscious Mac Users Are Turning On in 2026
- Public Wi-Fi on a Mac: The Real Risk Isn’t What You Think
- Mac Security Tools Compared: Firewalls, Antivirus and What Fits You
- Which Mac Apps Can You Trust? Signing, Notarization, Permissions and Network Behaviour
- Network Monitoring on a Mac: See Every Connection Your Apps Make
- Gatekeeper, XProtect and Notarization: What macOS Actually Catches — and Misses
- Mac Firewall vs Antivirus: You Need Both
- Ransomware on Mac: A Layered Defence That Actually Holds
- Spyware on a Mac: How to Spot It, What Removes It, How to Keep It Out
- The New Wave of Mac-Targeted Password-Stealing Malware
- Firewall AI for Mac: Setup and Optimization Guide
- Zero-Days on macOS: Why Unknown Threats Still Have to Phone Home
- How Chatty Is Your Mac? What macOS Says About You Without Asking
- Advanced Threat Detection on macOS: What the Layers Actually Do
- Yes, Macs Get Infected: A Short, Sourced History of Real Mac Malware
- Regain Control: How to Stop Data Harvesters Without Breaking Your Workflow
- Zero Trust at Home: Why Every App Should Earn Its Internet Access
- The Illusion of Built-In Security: Closing the Gaps the OS Leaves Wide Open
Privacy and online safety
Related: PrivateAI · University
- From Red Team to FireAI: A Note from Our CTO
- How to Tell Which App Is Using Your Mac’s Camera or Microphone Right Now
- Data Brokers: How Your Personal Information Leaks Even From a Well-Secured Mac
- File Metadata on a Mac: How to Find It and Remove It
- Ransomware Isn’t Just a Windows Problem: What Small Businesses on Mac Need to Know
- How Whistleblowers Protect Themselves: A Practical Digital Security Guide
- Securing a Mac for Remote Work: The Home-Office Checklist
- Mac Security for Creatives: Protecting Client Work Before It Ships
- The Hidden Risk on Roblox and Discord: A Parent’s Guide
- What Kevin Mitnick Can Still Teach You About Social Engineering
- The Data Broker Economy: Starving the Machine with Least Privilege
- The Enemy Inside: How "Trusted" Apps Sell Your Data Behind Your Back