Skip to content
NewA calmer, faster ZeroHour

Era Security dubs Greeter cloud campaign that leaves hello in logs

Original title: Attackers Say 'Hello' to a Different Scale of Cloud Attack

Dark Readingby Arielle Waldman ()8 min read
AI summary · x-ai/grok-4.7Summaries by Vani AI

About 90% of nearly 3,000 scanned Flowise deployments were unpatched against critical code-injection flaw CVE-2024-58351.

Era Security researchers, after a Shodan search found open Amazon DynamoDB instances, reported a cloud attack campaign they named Greeter. Roughly 90% of nearly 3,000 scanned Flowise deployments ran unpatched versions vulnerable to CVE-2024-58351. Honeypots, mostly on AWS, observed attackers mapping privileges, hunting credentials, using cloud metadata services, setting persistence, and leaving short hello messages in logs. The article text cuts off before naming victims or a fix.

Unconfirmed: The truncated report does not confirm CVE-2024-58351 was the entry point used against the honeypots or name real victims.

Key facts
  • Researchers started from publicly exposed DynamoDB instances that allowed download or deletion.
  • Flowise is an open-source platform for building LLM-powered AI agents.
  • Era Security said realistic honeypots can now be built in about two hours.
  • Attackers left benign-looking hello messages in logs to make detection harder.
Security assessment · AI, from the article

Impact: Code injection on exposed Flowise and access to cloud credentials

ProductAffected
Flowiseunpatched versions vulnerable to CVE-2024-58351
What to do
  1. 1.Identify internet-exposed Flowise deployments still vulnerable to CVE-2024-58351.
  2. 2.Restrict public access to Flowise and open Amazon DynamoDB instances.
Analyze CVE-2024-58351 with Vani AIImpact, affected versions, exploitation, mitigations and detection

Vulnerabilities mentioned

All
CVEVulnerabilityCVSSEPSSSignalsPublished
CVE-2024-58351
Flowise before 2.1.4 allows configuration to be injected into the Chainflow during execution via the overrideConfig option, supported in both the frontend web…
9.3<1%

Indicators of compromise

auto-extracted · verify before use · export allAll
TypeIndicatorContext
md57c184611bdd9df51982cc5bb2d5ba649vest-cloud-credentials#recipe=similar-items&sourcecontentid=7c184611bdd9df51982cc5bb2d5ba649) APT41 Delivers 'Zero-Detection' Backdoor to Harvest Cloud
md58f99d4e59eef3c1106085563b00071e8t-mythos-exploit-storm#recipe=similar-items&sourcecontentid=8f99d4e59eef3c1106085563b00071e8) CSA: CISOs Should Prepare for Post-Mythos Exploit Storm b
md5dedc2d87d0611657e649f3ccd3b03e96rastructure-crime-bots#recipe=similar-items&sourcecontentid=dedc2d87d0611657e649f3ccd3b03e96) TeamPCP Turns Cloud Infrastructure Into Crime Bots by Jai
md5e89675326c5a58c3d1742cab0ff9c2e8ts-power-offensive-ops#recipe=similar-items&sourcecontentid=e89675326c5a58c3d1742cab0ff9c2e8) Attackers Seize Exposed AI Endpoints to Power Offensive O

From the original article

8 min read at darkreading.com

Researchers dubbed an attack chain "Greeter" after discovering that attackers deliberately covered their trail by leaving "hello" messages in logs.

Dark Reading

By Arielle Waldman · darkreading.com

Read the full article at Dark Reading

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.