Astarot

[Astarot @egor65884]

@kriptoleidi If brain emulation becomes possible, the question of whether the copy is actually you will be just as important as the technology

[Astarot @egor65884]

@AmeliBesso8389 We’re used to thinking of this as a premonition of the future, but the idea of remembering a past cycle sounds incredibly interesting

[Astarot @egor65884]

Phone in someone else’s hands

After my phone has been in someone else’s hands, I don’t log into a single account until Claude Code finishes one pass.

The phone is back. I just don’t know yet what was done to it.

One prompt, 5 steps, read-only:

Profiles and management: looks for configuration profiles and MDM on iPhone, and apps with device admin rights on Android. Each of these takes a minute to install while your phone is out of your sight.

Accessibility: shows which apps can see your screen and read your notifications. That’s exactly how stalkerware reads your messages and SMS codes.

What got installed: flags anything installed outside the app store, anything recent, anything with a hidden icon or permissions a normal app has no reason to need.

The whole phone: runs Amnesty’s open-source toolkit (MVT) against known stalkerware indicators.

Report: one ranked table, from “all clear” to “your phone is compromised, change your passwords from another device.”

If the phone was with an ex or someone you’re afraid of, don’t wipe everything at once. A sudden cleanup can tip them off. Plan first, act second, and bring in a specialist if you need one.

It deletes nothing without your “yes,” never opens your messages or photos, and tells you straight up what it can’t see.

I’d rather give it 15 minutes than spend months wondering who’s reading my messages.

Full pass below, copy-paste ↓

[Astarot @egor65884]

@ViktoriaLee10 What if both are genuinely convinced they’re the original? Neither could prove it from memory alone

[Astarot @egor65884]

@Yarilo7brigada One person, with properly configured agents, can now do what previously required several specialists

[Astarot @egor65884]

@SSSvinosvin The original “infinite money glitch” that turned out to be infinite debt 💀

[Astarot @egor65884]

@Damir_Akaza The story sounds like a techno‑thriller. I would first check whether there are real logs and independent confirmation of the cause of the infection

[Astarot @egor65884]

Agreed, and that's why the last section is "what I can't see" rather than a green checkmark. The check gives you evidence and a shortlist, not a clean bill of health - it can't see firmware, a hardware implant, or someone who just knows your passcode. Knowing where the blind spots are is the real output. "Peace of mind" that ignores them is worse than nothing

[Astarot @egor65884]

Phone in someone else’s hands

After my phone has been in someone else’s hands, I don’t log into a single account until Claude Code finishes one pass.

The phone is back. I just don’t know yet what was done to it.

One prompt, 5 steps, read-only:

Profiles and management: looks for configuration profiles and MDM on iPhone, and apps with device admin rights on Android. Each of these takes a minute to install while your phone is out of your sight.

Accessibility: shows which apps can see your screen and read your notifications. That’s exactly how stalkerware reads your messages and SMS codes.

What got installed: flags anything installed outside the app store, anything recent, anything with a hidden icon or permissions a normal app has no reason to need.

The whole phone: runs Amnesty’s open-source toolkit (MVT) against known stalkerware indicators.

Report: one ranked table, from “all clear” to “your phone is compromised, change your passwords from another device.”

If the phone was with an ex or someone you’re afraid of, don’t wipe everything at once. A sudden cleanup can tip them off. Plan first, act second, and bring in a specialist if you need one.

It deletes nothing without your “yes,” never opens your messages or photos, and tells you straight up what it can’t see.

I’d rather give it 15 minutes than spend months wondering who’s reading my messages.

Full pass below, copy-paste ↓

[Damir Akaza @Damir_Akaza]

My friend checked into a hotel with his girlfriend, ran a single script in Claude Code, and found out in 90 seconds that a hidden camera was running right behind the bathroom mirror

After what he told me, neither of us even unpacks a suitcase in a hotel or an Airbnb anymore until a laptop by the door finishes one full sweep

One prompt runs a 5-stage audit strictly in read-only mode:

-> Wi-Fi radar: scans the local network via nmap, resolves manufacturers from MAC addresses, and instantly flags hidden cameras and ports streaming continuous video

-> Bluetooth interceptor: listens across the room via the bleak library for unknown AirTag, Tile, and SmartTag trackers while spotting hotel locks on ancient firmware that open from the hallway

-> Spyware detector: audits your phone against the open-source Amnesty MVT toolkit and exposes hidden iconless processes quietly leaking your location or mic audio

-> Optical lens sweep: if the hotel isolates guest devices on the network, the agent walks you step by step with a flashlight across every bathroom mirror, vent, and angle facing the bed

-> Priority report: builds one clean table ranking the risk level of every suspicious device and tells you the exact next move: from demanding another room to calling the police

The agent never logs into anything, never hacks foreign devices, and translates 500 lines of raw network traffic into plain English in 2 minutes

I would rather spend 2 minutes with an open terminal in the hallway than spend the whole trip being watched

The full sweep prompt is below, just copy and paste ↓

Bookmark this post before your next check-in

[Astarot @egor65884]

@ViktoriaLee10 Yeah, that's the whole point of doing it before you log in - a profile or an accessibility app that's already there will just scoop up the first password you type. Check first, log in second. Order's free; redoing it isn't

[Astarot @egor65884]

@AmeliBesso8389 and it's the part people get wrong under stress. If an abuser is the one watching, a sudden wipe tells them you know, and that's often the most dangerous moment. Document first, keep a safe device for passwords, talk to a specialist

[Ackerman @Yarilo7brigada]

this Grok Bot terminal claims it caught 37 fake “done”s from its own bots in a week

wild result. but the six-agent setup is what’s worth looking at

instead of taking a bot’s word for it, the author split the process:

→ a dispatcher assigns tasks and attaches context from the brain
→ a worker does the job
→ an evidence agent collects command output and files
→ a verifier only reads and gives a verdict
→ a handoff agent writes the note for the next bot
→ a logging agent records decisions and the reasons behind them

the terminal puts the whole process on one screen
a task board next to the activity feed
accept / reject / retry / escalate verdicts
evidence checks before a task closes
which brain files the bot read before starting
stats on accepted and rejected tasks

the author reports:
tasks this week: 286
rejected by the verifier: 37
accepted on first try: 81.4%
events logged: 1,964
these are self-reported numbers, not an independently verified audit

and 81.4% accepted on first try doesn’t settle anything. the verifier might just be too soft

what makes the architecture interesting is the ability to trace a task backwards

who assigned it
what context was passed
what evidence was collected
why the verdict was what it was
who picked it up next

that’s far more useful than a green task board with no history behind it

if you’re building something like this, make verification transparent before the “done” rate gets impressive

save this for splitting the worker from the checker, not for the number of closed tasks

[Astarot @egor65884]

@kriptoleidi The copy might be convinced it survived. That doesn't answer what happened to the original

[Astarot @egor65884]

@AmeliBesso8389 This isn’t just a flight, it’s a transition to a completely different state

[Nezuko @Nezukoa4]

He was just over twenty when he realized the $1.5 million was simply gone.

He didn’t spend it. It was taken by the man he had given full control of his money.

Late ’90s. Kenan Thompson finishes his run on Nickelodeon - All That and Kenan & Kel. Years of work had added up to a check of roughly $1.5 million. For a young actor, that was life-changing money.

Instead of learning how to manage it himself, he (still a teenager) signed a power of attorney for an accountant. The man promised to “handle everything.” The family agreed. The red flags were ignored.

Later Kenan decided to buy a house in Atlanta. At closing, the accountant was supposed to show up with the money. He never came. The phone stayed silent. The money turned out to be a ghost.

$1.5 million - the entire Nickelodeon payday - vanished.

He had to file for bankruptcy. It took years to rebuild his credit. All while he kept working and making people laugh on camera.

His first paycheck ever was $800 for a fried chicken commercial. He was about twelve. At the time it felt like ten thousand dollars.

Today Kenan is the longest-running cast member in Saturday Night Live history. He still watches his money himself and says one simple thing: “I never give anyone power of attorney. Ever.”

The lesson: when you hand over control of your money, you’re not just handing over the account. You’re handing over the chance to know what’s happening until it’s too late.

In the interview he talks about sitting in a car outside that man’s house waiting, and what changed after. Watch him, then decide.

The most expensive mistake of youth, or the most useful lesson about trust?

[Astarot @egor65884]

@ViktoriaLee10 It seems to me that the key issue is not memory, but the continuity of subjective experience

[Astarot @egor65884]

Anthropic engineer:

“Anthropic pays engineers $350–950k because harnesses still decide whether an agent survives in production or not”

“I’m a product engineer - I build agent harnesses and roll them out with big customers. The goal is to make sure it doesn’t fall apart”

The Anthropic engineer showed how the interfaces for building agents evolved: from the plain Messages API to the Claude Agent SDK and then to Managed Agents, where the loop already runs on the company’s side.

Opus → Agents → Harness → Server-side loop

In the talk they broke down the key changes:

step 1 → on the Messages API the entire loop, context, and tools stay with the developer
step 2 → the Agent SDK gives you the harness and tools, but you still have to host and scale it yourself
step 3 → with Managed Agents the loop runs on the server, brain and hands are separated, sessions are persistent
step 4 → after a crash you can resume from the log instead of starting over
step 5 → past sessions go into memory and a process Anthropic calls “dreaming”

Most teams still spend weeks building this infrastructure by hand.

With Managed Agents the platform takes some of that work.

Save and watch.
Then read the full harness engineering guide below ↓

[Astarot @egor65884]

@Damir_Akaza Ollama plus a local knowledge base is a solid starting point for private document search. The hardware is the real trade-off

[Damir Akaza @Damir_Akaza]

It is becoming f*cking dangerous

When Anthropic opened up local configuration for Claude Code, they had no idea users would completely cut off their cloud servers and build a 100% offline data bunker for $0

-> Set up a folder called LOCAL VAULT on your computer

-> Download the free Ollama app and run a local model directly on your hardware so all text processing happens without external servers

-> Open the Claude Code terminal inside that folder and configure local API routing so the agent runs exclusively through your home model without sending requests to the cloud

-> Make a new subfolder named PRIVATE MEMORY inside the root directory to store all your sensitive financials, contracts, and personal records

-> Add a CLAUDE.md file to the root folder with strict instructions to log all critical context inside the PRIVATE MEMORY folder and block any outbound network connections while processing documents

-> Install the open-source Khoj or LocalGPT tool from GitHub to connect all your PDF files, spreadsheets, and notes into a unified local knowledge base right on your hard drive

-> Bundle this entire chain via skills into a single local command that launches your private assistant with one click even when your Wi-Fi is turned off

The entire setup takes roughly 20 minutes

Get your own fully local second brain with persistent memory and instant search across all your documents and finances that lives exclusively on your hard drive and physically cannot leak your data to the internet

Bookmark this post if you value your privacy and refuse to feed corporations with your personal documents

[Astarot @egor65884]

@AmeliBesso8389 This moment with Histoire’s promise always gets to me. Live your own life, not hide behind someone else’s name