Real Agent Cases

Check a 'take-home task' repo from a job offer for malware with ChatGPT before running it

Posted
by @makowskid
Last checked
Method:
How we verify

Evidence: Described by posterThe person describes what happened in words, without screenshots or video.

Outcome

A developer offered a 'unicorn' job taking over a software project was asked to review and run the company's GitHub repository. Instead of cloning or running it, he asked ChatGPT to analyze it, and says it found hidden malware in what looked like a harmless library file, built to connect to an outside server and steal files, keys and credentials when the project was started. He calls it a common scam aimed at job seekers.

Agent

ChatGPT

Prompt Recreated prompt

The source didn't share its exact prompt, so we wrote this from what it describes.

Run this task now, exactly as written below. Everything you need is here, so start straight away and only stop if something required is missing or a rule below tells you to.

Goal: Check a job offer and any files or code they sent me for scam warning signs before I open or run anything.

Fill in before running:
- The job message and company name: [PASTE]
- Links or files they sent (don't run them): [LINKS]
- What they're asking me to do: [DETAILS]

Rules:
- List red flags (urgency, payment requests, odd links, requests to run code or install software).
- For code, look for hidden network calls, obfuscated text, or scripts that run on install or start.
- Check whether the company and recruiter appear to be real, citing sources.
- Tell me plainly: safe to proceed, proceed with caution, or likely scam.
- If a site asks for a login, a CAPTCHA or payment details, stop and hand control back to me.
- Ask for my OK before you pay, book, send, submit or cancel anything.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.

Output: A red-flag list and a clear verdict with reasons.

You'll know it worked when: I make a decision without having run or installed anything from them.

Try this in:

This prompt has blanks to fill in, like [PASTE]. We'll copy it instead of filling it in, so you can edit it before you send.

Opens in a new tab. We don't see what you send.

These links are only a shortcut. We're not affiliated with OpenAI, Google or xAI, and your results may not match the source.

Optional stronger version

Our rewrite with extra checks. The poster didn't test it.

Run this task now, exactly as written below. Everything you need is here, so start straight away and only stop if something required is missing or a rule below tells you to.

Goal: Check a job offer and any files or code they sent me for scam warning signs before I open or run anything.

Fill in before running:
- The job message and company name: [PASTE]
- Links or files they sent (don't run them): [LINKS]
- What they're asking me to do: [DETAILS]

Rules:
- List red flags (urgency, payment requests, odd links, requests to run code or install software).
- For code, look for hidden network calls, obfuscated text, or scripts that run on install or start.
- Check whether the company and recruiter appear to be real, citing sources.
- Tell me plainly: safe to proceed, proceed with caution, or likely scam.
- If a site asks for a login, a CAPTCHA or payment details, stop and hand control back to me.
- Ask for my OK before you pay, book, send, submit or cancel anything.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.
- Never run code or install apps from a stranger on your main computer; report scam listings to the platform.

Output: A red-flag list and a clear verdict with reasons.

You'll know it worked when: I make a decision without having run or installed anything from them.

Before you finish: re-read my goal and constraints, check your answer against each one, and list anything you assumed or couldn't confirm.

Source

https://x.com/makowskid/status/2010206469069709692

  • Spot a fake IRS letter as a scam by checking it with ChatGPTRecreated promptEvidence: Described by posterBeginner-friendly

    When a letter claiming to be from the IRS arrived, Sara Wilbur's mother suspected a scam, checked it with ChatGPT and was right. Wilbur, who had encouraged her to trust her instincts, watch for red flags and use ChatGPT, shared it as an example of fraud awareness working.

    ChatGPT · Personal · Source date:

  • Rathesh Prabakar, a senior software engineer in Bengaluru, quit in October 2025 without a job lined up. Prabakar applied before work and prepared after, and asked ChatGPT to act as an experienced résumé writer. For each interview, Prabakar gave it their skills, experience and the job description plus the time available and asked for a prep plan; a high share of the likely questions it suggested came up. Questions Prabakar couldn't answer went back to ChatGPT afterwards, and doubtful answers were cross-checked on Google. With a referral from a LinkedIn contact, Prabakar accepted an offer in early November.

    ChatGPT · Personal · Source date:

  • On their way to interviews at Microsoft, Bain and Sweetgreen, a job seeker who writes the Ambitious Slacker newsletter pasted job descriptions into ChatGPT and had it interview them in voice mode, answering out loud and asking for feedback. That practice helped. Having ChatGPT write answers to memorise backfired: after a final round at Target, the interviewer said they were the best candidate but the company chose an internal candidate because they sounded too scripted. Once they stopped memorising and practised a few stories out loud, they did much better.

    ChatGPT · Personal · Source date: