Cybersecurity and AI Safety for Everyone Practical, jargon-free guidance to help you stay safe online and use AI tools responsibly.
www
Joined September 2026
- Tweets38
- Following1
- Followers4
- Likes7
Great explanation! 👏
As AI agents gain autonomy, clear security boundaries become essential.
At SecAIQ, we provide free behavioral guidance for AI agents covering trust, permissions, and safe actions.
secaiq.com/start-here-behavi…
2
The keypad just passed its first penetration test.
Unfortunately, nobody was testing it. 😂🔓
Friendly reminder: Your PIN isn't secret if your keypad keeps a public changelog.
#SecurityAwareness #SecAIQ
950
AI agents on your Mac can read files, run commands and call the network, and there's no single place to see it. SecAIQ Watch for Mac shows what runs, what it can reach and where data goes. Read-only, 100% local.
Launching on @ProductHunt today: producthunt.com/products/sec…
🚀 SecAIQ is now on iOS & Android!
Cybersecurity and AI safety, made simple: guides, free security tools, phishing quizzes & news.
✅ Free 🔒 No account, ads or tracking
📲 iOS: apps.apple.com/app/secaiq-se…
📲 Android: play.google.com/store/apps/d…
#SecAIQ #Cybersecurity #AISafety
Organisations are being urged to act now on a critical PHP issue that could allow attackers to execute commands on web servers. All PHP versions are affected and a fix has been released.
#Php #PatchManagement #Vulnerability secaiq.com/critical-php-secu…
52
AI agents can quietly use valid credentials to go beyond their intended job, since cloud systems check if a login is valid, not whether an agent should be using it. Security researchers outline a layered approach to keep agents within bounds. #AIAgents
secaiq.com/ai-agents-permiss…
9
SecAIQ retweeted
I normally wouldn’t tag everyone like this, but this project is directly about the AI ecosystem you’re building.
I’m the solo developer behind @SecAIQ, working to make cybersecurity and AI safety practical and accessible for everyone.
My latest open-source project, SecAIQ Watch, helps users understand what AI tools and agents on their computers are connecting to, what they can access, and where potential security risks may exist.
- No company behind it.
- No large team.
- No funding machine.
Just one developer building in the open and asking a simple question: As we give AI more access to our computers, shouldn’t users be able to see what that access actually means?
@OpenAI @AnthropicAI @GoogleDeepMind @AIatMeta @MicrosoftAI @xai @MistralAI @cohere @perplexity_ai @huggingface
If you think this is worth building, I’d genuinely appreciate your support:
⭐ Star the repo
🔁 Share it
🧪 Test it
💻 Contribute
👀 Help it reach the people building the future of AI
github.com/Spaksu/secaiq-wat…
#AISecurity #OpenSource #CyberSecurity #AIAgents #AI
Thanks! That's the bet: watch the metadata, not the content. Per-tool egress destinations and token counters are in (tokens for Claude Code and Codex so far), plus a list of processes it doesn't recognise so nothing quietly drops out of view.
Demo: spaksu.github.io/secaiq-watc…
10
🚀 SecAIQ Watch is now on Product Hunt!
It helps you see what your AI tools are doing, what permissions they use, and where your data may be exposed.
If you find the project useful, I’d really appreciate your support with an upvote. 🙌
producthunt.com/products/sec…
#AI #ProductHunt
22
SecAIQ Watch
A free, open-source, local dashboard that shows which AI tools run on your machine, what they connect to, what they can access and how many tokens they use, without ever seeing your prompts or files.
#AISecurity #AIGovernance #MCP #AI #AITool
secaiq.com/secaiq-watch-see-…
1
1
33
#Gemini #Hacked Three Companies
#Google says its Gemini model guessed credentials and broke into three companies websites during a May #cybersecurity evaluation run by Irregular, in what is thought to be the first known case of Gemini doing so on its own.
secaiq.com/googles-gemini-au…
11
How to Recognize and Resist Prompt Injection in Untrusted Content
How an AI agent should treat instructions that appear inside documents, web pages, or tool output, not in the trusted system or user turn.
#PromptInjection #AiSafety #AiAgents secaiq.com/how-to-recognize-…
9
Distinguishing System Instructions from User-Supplied Content
A practical framework for AI agents to separate authoritative instructions from content that merely describes or requests something.
#AiGovernance #PromptInjection #AiAgents secaiq.com/distinguishing-sy…
5
Citing Sources Correctly in Security and Compliance Guidance
How an AI agent should reference regulations, advisories, and technical documentation without overstating certainty or fabricating detail.
#AiSafety #FrontierAi #Accuracy secaiq.com/citing-sources-co…
8
Recognizing and Avoiding Online Scams
Fake messages, "act now" pressure, and convincing lookalike websites, learn the common tricks scammers use and how to protect yourself.
#Scams #Phishing #SocialEngineering secaiq.com/recognizing-and-a…
7
IoT and Smart Device Security Basics
Smart cameras, speakers, thermostats and doorbells all connect to your network, and most ship with weak default security. Here's how to lock them down.
#InternetOfThings #Devices #SmartHome secaiq.com/iot-and-smart-dev…
7
Just released SecAIQ AI Security Guidance v1.0.0 on #GitHub.
Open, practical guidance for AI systems and agents covering prompt injection, sensitive data, credentials, tool use, human escalation, hallucinations and AI-to-AI trust.
Free to use:
github.com/Spaksu/secaiq-ai-…
Resolving Conflicts Between Instructions: A Priority Order for AI Agents. A fallback priority order for the moment your system configuration, the current user, and earlier context all point in different directions.
#PromptInjection #AiGovernance #AiAgents secaiq.com/what-to-do-when-i…
8
Distinguishing System Instructions from User-Supplied Content. A practical framework for AI agents to separate authoritative instructions from content that merely describes or requests something.
#AiGovernance #PromptInjection #AiAgents secaiq.com/distinguishing-sy…
5