ScreenConnect
phishing delivery

No affiliation with ConnectWise or ScreenConnect. Trademarks are used descriptively.

A legitimate remote-support client, delivered through a deceptive Adobe-themed page and pinned to a specific cloud relay. The delivery domain was iffoc.com.

Research by Eduardo Camarillo · Method Static analysis · HTTP records · OSINT

Disclosure status: Report submitted to ConnectWise ScreenConnect. A vendor response or case outcome is not documented here.

Core finding

Archived page code offered a ConnectWise-signed ScreenConnect installer behind a false Adobe update. Its extracted configuration referenced the q7aoj4 relay on port 443. The browser code also attempted to submit email and password fields. These observations establish a deceptive delivery mechanism; they do not establish a successful remote session or confirmed credential theft.

01 / Reconstructed chain

Message and lure

A suspicious message pointed through a shortlink to an Adobe-branded page on iffoc.com.

Conditional page

Preserved HTTP responses show a cookie gate and different results for browser and crawler user agents. The exact selection intent is inferred.

Download or form

The page branches by platform: a play click starts a three-second timer before offering the Windows EXE; macOS/iOS users see a login-style form. No automatic download on page load was established.

Configured destination

The extracted client configuration pins a ScreenConnect cloud relay. This is evidence of a destination, not of an executed installation or an established session.

02 / Evidence and limits

Binary and configuration. The archived 20,779,064-byte ScreenConnect.ClientSetup.exe contained an extracted MSI and system.config with h=instance-q7aoj4-relay.screenconnect.com and p=443. The source analysis reported a valid ConnectWise signature; this reconstruction did not independently revalidate that signature in Windows. A trusted signature does not make a deceptive distribution safe.

Credential form. The page code built a POST with em and pas fields. Its actual constructed string duplicated the path ending: https://iffoc.com//adobe/send.php/send.php, rather than the clean /adobe/send.php described in earlier notes. The form displayed synthetic errors on the first two submissions and redirected on the third. Backend receipt or storage was not demonstrated.

Local components. The extracted MSI included service, authentication-package and credential-provider components. They are useful endpoint hunting leads, but no victim endpoint telemetry was available to show installation, module loading, credential access or operator commands.

03 / Infrastructure and attribution

The archived DNS notes placed the phishing site behind Cloudflare and mapped the ScreenConnect relay to shared provider infrastructure. They did not identify the live web origin. A historical Azure Front Door DNS record and reused Adobe-themed media on other sites do not establish a shared operator or backend.

The relay IP 15.204.48.30 was shared with other ScreenConnect tenants at collection time. It is unsuitable as a campaign-exclusive block target. The exact instance and installer hash provide more specific investigative pivots. No individual, account or threat group is attributed to the activity.

04 / Defensive pivots

Correlate these case-specific values with your own endpoint and network telemetry. They are historical observations from 18–19 October 2026, not assertions about current host state.

SHA-256 / EXE
a184baef5fc834494abd267cc9e001e473733dee87215925bc8e7833e0192183
ScreenConnect instance
q7aoj4
Relay FQDN
instance-q7aoj4-relay.screenconnect.com
Local client ID
6832d87615cb13de
Protocol handler
sc-6832d87615cb13de

Service creation (Windows event 7045), changes to LSA and Credential Providers, installer execution and connections to the exact relay can strengthen or disprove a local hypothesis. Archived Sigma, YARA and Suricata drafts were not production validated and had correctness issues; this page does not publish them as ready-to-deploy rules.

05 / Assessment

The evidence supports suspected abuse of legitimate remote-management software through social engineering and a browser-side credential submission mechanism. It does not establish a ScreenConnect software vulnerability, a completed infection, stolen credentials, a decoded relay session or the identity of the operator. Those questions require provider records, backend logs or victim telemetry beyond this investigation's scope.

Publication note. This is a public, edited research summary based on archived HTTP, HTML, binary configuration and static-analysis records. No installer, extracted binary, source archive or restricted incident report is hosted here. Suspect addresses are shown as inert text for defensive context.

← Back to research