# CISA KEV: Active Exploitation of Critical RCE Flaws in Synology DSM (CVE-2024-10441), Apache Axis (CVE-2026-90436), REDCap (CVE-2026-90817) and LiteLLM (CVE-2026-40217)

> CISA added four vulnerabilities to the KEV catalog on 2026-11-18 citing evidence of active exploitation: Synology DSM/BeeStation OS, Apache Axis 1.x through 1.4, Vanderbilt REDCap from 13.3.0, and the BerriAI LiteLLM proxy. CISA did not attribute the activity to a single campaign; separate reporting links recent Synology NAS attacks to C3RB3R ransomware and Rapid7 observed CVE-2026-90436 exploited against Confluence Data Center and N-able N-central.

- **Published:** 2026-11-19T11:59:12Z
- **Last reviewed:** 2026-11-19T11:59:12Z
- **Canonical:** https://intel.threadlinqs.com/threat/TL-2026-6450
- **Card image:** https://intel.threadlinqs.com/og/threat/TL-2026-6450.png
- **ID:** TL-2026-6450
- **Severity:** CRITICAL (CVSS 9.8)
- **Category:** VULNERABILITY
- **Status:** ACTIVE
- **Detections:** 9 (1 shown free below) · **IOCs:** 25 (3 shown free below)
- **CVEs:** CVE-2024-10441, CVE-2026-90436, CVE-2026-90817, CVE-2026-40217, CVE-2026-47101, CVE-2026-47102

## Description

On 2026-11-18 CISA added four CVEs to the Known Exploited Vulnerabilities catalog: CVE-2024-10441 (Synology DiskStation Manager, 'Improper Encoding or Escaping of Output'), CVE-2026-40217 (BerriAI LiteLLM, 'Unprotected Alternate Channel'), CVE-2026-90436 (Apache Axis, 'Unsafe Reflection') and CVE-2026-90817 (Vanderbilt REDCap, 'Code Injection'). The CISA alert gives no technical detail, CWE IDs or campaign attribution, and the four listings are unrelated to each other beyond the KEV addition. Per GBHackers, federal (FCEB) remediation deadlines under BOD 26-04 are 2026-11-21 for the Synology, Axis and REDCap entries and 2026-12-02 for LiteLLM.

Synology DSM (CVE-2024-10441, CVSS 9.8): an output-escaping flaw in the system plugin daemon allows unauthenticated remote code execution. It was fixed in DSM 7.1.1-42962-7, 7.2-64570-4, 7.2.1-69057-6 and 7.2.2-72806-1 (and 6.2.4-25556-8 for DSM 6.2 per The Hacker News); BeeStation OS is affected before 1.1-65374. Synology's advisory SA_24_20 also covers CVE-2024-50629 (limited file read) and CVE-2024-10445 (limited file write). In late October 2026 Synology warned of ongoing C3RB3R ransomware attacks against NAS devices running outdated DSM; C3RB3R encrypts files with the .L0CK3D extension, drops read-me3.txt ransom notes and skips 27 directories and extensions such as BAT, DLL, EXE, LNK, MOD, MSI and SYS. Synology did not name the exploited CVE, and Qihoo 360 reportedly tied C3RB3R to two different DSM flaws fixed in September 2026 (CVE-2026-13684 and CVE-2026-13639, SA_26_13), so the link between C3RB3R and CVE-2024-10441 is not confirmed by the vendor.

Apache Axis (CVE-2026-90436, CVSS 9.8, CWE-470): disclosed 2026-11-06, an unauthenticated unsafe-reflection RCE in the end-of-life Axis 1.x line (through 1.4). The SOAP engine processes certain messages before service-level authentication, so application auth does not protect exposed endpoints. No fixed Axis 1.x release is planned; Apache recommends migrating to Axis2/Java or applying the source patch (commit 41f3b4e8d04e310ca81f86b92630ab2c11d608bd). Rapid7 MDR detected exploitation against a customer's Atlassian Confluence Data Center on 2026-11-13 via the legacy Remote API (XML-RPC & SOAP) endpoint /plugins/servlet/soap-axis1/confluenceservice-v2 (non-default; fixed in 9.2.28, removed in 10.x). Unauthenticated POSTs returned HTTP 500 yet the Confluence Java process spawned Bash and curl as the confluence account; the operator then ran hundreds of hex-encoded commands and scripts pulled from Pinggy and Cloudflare Tunnel hosts, enumerated the host, read encryption keys and database settings, logged into MySQL with Confluence's own credentials, harvested page bodies and attachments (including a Luhn-validated search for payment card numbers), queried Atlassian Crowd, attempted AS-REP roasting with Impacket GetNPUsers.py (failed), queried a cloud data warehouse about four hours in, and signed in to Microsoft 365 with a service account via legacy authentication about twelve hours in. A proof-of-concept targeting N-able N-central was published 2026-11-12 and N-able reported exploitation; N-central is vulnerable when the Legacy API is enabled (including 2026.4 HF2) and fixed in 2026.4 HF3 (2026.4.1.39). A Metasploit module for Confluence was released 2026-11-17.

REDCap (CVE-2026-90817, CVSS 9.8): a code-injection flaw in public survey routing (survey passthrough) and Data Import processing, affecting REDCap from 13.3.0; unaffected versions are 16.0.49, 17.3.10 and 17.4.4. No account is needed but the attacker must know a valid public survey hash. NVD published the CVE on 2026-09-20. A public Nuclei template exists. Specific in-the-wild exploitation details were not available in the sources reviewed; a Unit 42 link to an XMRig campaign and a separate UNC6508 REDCap campaign were mentioned by secondary sources but could not be confirmed as tied to this CVE.

LiteLLM (CVE-2026-40217, CVSS 8.8, CWE-420): the custom-code guardrail test endpoint POST /guardrails/test_custom_code executes submitted Python behind a keyword deny-list rather than real isolation. Affected: 1.81.8 up to but not including 1.83.10 (fixed 2026-04-15, disclosed 2026-04-10). Zenity Labs observed 569 POSTs between 2026-04-15 and 2026-05-24 from eight source IPs, guessing default master keys (sk-1234, sk-litellm-master-key, sk-admin, sk-changeme), confirming code execution with a {'probe':1} marker, bypassing the deny-list by rebuilding dunder names from string fragments and recovering the real __import__ via generator/coroutine internals, then dumping os.environ for API keys and secrets or running shell commands; the official Docker image runs as root. Public PoCs exist (Obsidian Security / X41 chain CVE-2026-47101 -> CVE-2026-47102 -> CVE-2026-40217).

Defender priorities: patch or disable the exposed component, hunt for web-server or application processes spawning shells, review outbound tunnel-service traffic, rotate all credentials and API keys stored on or reachable from affected hosts, and preserve evidence before patching because patching does not remove a prior intrusion.

## MITRE ATT&CK

- T1190 Exploit Public-Facing Application
- T1078.004 Cloud Accounts
- T1059.004 Unix Shell
- T1059.006 Python
- T1027 Obfuscated Files or Information
- T1110.001 Password Guessing
- T1552 Unsecured Credentials
- T1558.004 AS-REP Roasting
- T1082 System Information Discovery
- T1213 Data from Information Repositories
- T1005 Data from Local System
- T1572 Protocol Tunneling

## Free detection (Splunk SPL, Microsoft KQL, Sigma)

**Apache Axis 1.x SOAP Unsafe Reflection Exploitation (CVE-2026-90436) — Exploitation Phase** (`TL-2026-6450-SPL-001`)

Detects unauthenticated HTTP POSTs to Axis 1.x SOAP servlet endpoints that return HTTP 500, the signature Rapid7 observed against Confluence Data Center (/plugins/servlet/soap-axis1/confluenceservice-v2) during CVE-2026-90436 exploitation (also applicable to N-able N-central Legacy API). The SOAP engine processes crafted messages before service authentication, so the server errors out while the payload still executes. A hit means a host is being actively probed or exploited with the Metasploit module / public PoC.

- **Languages:** Splunk SPL, Microsoft KQL, Sigma
- **ATT&CK:** T1190
- **Data sources:** web_access_logs, index=web sourcetype=access_combined, index=web sourcetype=iis
- **Severity:** critical
- **Confidence:** high

### Splunk SPL

```spl
index=web (sourcetype=access_combined OR sourcetype=iis OR sourcetype="nginx:plus:access") ("/plugins/servlet/soap-axis1/" OR "confluenceservice-v2" OR "/axis/services/" OR "/services/AdminService") earliest=-24h
| eval src=coalesce(src,clientip,c_ip), dest=coalesce(dest,host), http_method=upper(coalesce(method,http_method,cs_method)), url=coalesce(uri_path,cs_uri_stem,uri), http_status=coalesce(status,sc_status)
| where http_method="POST"
| stats count AS requests sum(eval(if(http_status=500,1,0))) AS http_500 values(url) AS urls values(http_status) AS statuses min(_time) AS first_seen max(_time) AS last_seen BY src dest
| where http_500>=1
| convert ctime(first_seen) ctime(last_seen)
```

### Microsoft KQL

```kql
_Im_WebSession(starttime=ago(24h), endtime=now())
| where HttpRequestMethod =~ 'POST'
| where Url has_any ('/plugins/servlet/soap-axis1/', 'confluenceservice-v2', '/axis/services/', '/services/AdminService')
| summarize Requests=count(), Http500=countif(EventResultDetails == '500'), Urls=make_set(Url, 10), FirstSeen=min(TimeGenerated), LastSeen=max(TimeGenerated) by SrcIpAddr, DstIpAddr
| where Http500 >= 1
| project FirstSeen, LastSeen, SrcIpAddr, DstIpAddr, Requests, Http500, Urls
```

### Sigma

```yaml
title: Apache Axis 1.x SOAP Unsafe Reflection Exploitation Attempt (CVE-2026-90436)
id: 6f1c0a52-3b7d-4e19-8a64-2d9e5c7b1a01
status: experimental
description: Detects unauthenticated POST requests to Axis 1.x SOAP servlet endpoints that return HTTP 500, as seen in CVE-2026-90436 exploitation against Confluence Data Center and N-able N-central.
references:
  - https://www.cisa.gov/news-events/alerts/2026/11/18/cisa-adds-four-known-exploited-vulnerabilities-catalog
  - https://rapid7.com/blog/post/etr-rapid7-detected-exploitation-of-cve-2026-90436-against-atlassian-confluence-data-center-n-able-n-central-also-exploited
author: Threadlinqs Team
date: 2026-11-19
tags:
  - attack.initial_access
  - attack.t1190
  - cve.2026.90436
logsource:
  category: webserver
detection:
  selection_uri:
    c-uri|contains:
      - '/plugins/servlet/soap-axis1/'
      - 'confluenceservice-v2'
      - '/axis/services/'
      - '/services/AdminService'
  selection_method:
    cs-method: 'POST'
  selection_status:
    sc-status: 500
  condition: selection_uri and selection_method and selection_status
falsepositives:
  - Legitimate internal SOAP clients calling an Axis 1.x service that fails with a server error
  - Vulnerability scanners run by the security team
level: high
```

**Known false positives:**

- Legitimate internal SOAP clients calling an Axis 1.x service that returns HTTP 500
- Authorized vulnerability scans against Confluence or N-central

**Tuning notes:** Exclude known scanner and internal integration source IPs. The 500 filter reflects Rapid7's observation; if your reverse proxy rewrites errors, also alert on any external POST to the soap-axis1 path. If Confluence Remote API (XML-RPC & SOAP) is disabled, any hit is high-fidelity. Extend the URI list with other Axis 1.x service paths exposed in your environment.

### 8 more detection rules (Splunk SPL, Microsoft KQL, Sigma), unlock with TL Creds

- LiteLLM Guardrail Test Endpoint Abuse (CVE-2026-40217) — Exploitation Phase (locked: unlock with TL Creds)
- Confluence Credential, Config and Attachment Harvesting plus Host Discovery — Post-Exploitation (locked: unlock with TL Creds)
- Contact with Reported CVE-2026-90436 and LiteLLM Exploitation Infrastructure — C2 (locked: unlock with TL Creds)
- C3RB3R Ransomware File Artifacts ([indicator] and [indicator]) — Impact (locked: unlock with TL Creds)
- Hex-Decoded Command Piped to Shell and Impacket GetNPUsers AS-REP Roasting — Post-Exploitation (locked: unlock with TL Creds)
- Microsoft 365 Legacy-Authentication Sign-In by Service Account / Reported Actor IP — C2 / Lateral Movement (locked: unlock with TL Creds)
- Web/Application Process Spawning Shell or Download Utility — Post-Exploitation (locked: unlock with TL Creds)
- Server Egress to Pinggy, Cloudflare Tunnel and [indicator] — C2 / Tooling Staging (locked: unlock with TL Creds)

Sign up free to unlock the detection rules: https://threadlinqs.com/?modal=signup

## Sample indicators (3 of 25)

Defanged, TLP:CLEAR.

- IP address: `192.238.253[.]176`
- Domain: `wpefh-192-238-253-176[.]free[.]pinggy[.]net`
- IP address: `5.133.47[.]243`

Full set of 25 indicators unlocks with TL Creds: https://threadlinqs.com/tl-creds

## Sources

- [CISA Adds Four Known Exploited Vulnerabilities to Catalog](https://www.cisa.gov/news-events/alerts/2026/11/18/cisa-adds-four-known-exploited-vulnerabilities-catalog)
- [Hackers Exploit Critical RCE Vulnerabilities in Synology NAS, Apache Axis and REDCap, CISA Warns (GBHackers)](https://gbhackers.com/critical-rce-vulnerabilities/)
- [Rapid7: Detected Exploitation of CVE-2026-90436 against Atlassian Confluence Data Center, N-able N-central also Exploited](https://rapid7.com/blog/post/etr-rapid7-detected-exploitation-of-cve-2026-90436-against-atlassian-confluence-data-center-n-able-n-central-also-exploited)
- [Zenity Labs: Threat Actors Are Trying to Use LiteLLM's Guardrail Tester to Run Code as Root](https://labs.zenity.io/p/threat-actors-are-trying-to-use-litellm-s-guardrail-tester-to-run-code-as-root)
- [The Hacker News: Synology NAS Devices Targeted in C3RB3R Ransomware Attacks](https://thehackernews.com/2026/11/synology-nas-devices-targeted-in.html)
- [BleepingComputer: Synology warns customers of ongoing C3RB3R ransomware attacks](https://www.bleepingcomputer.com/news/security/synology-warns-customers-of-ongoing-c3rb3r-ransomware-attacks/)
- [Obsidian Security: Breaking LiteLLM - From Low-Privilege User to Admin and RCE (CVE-2026-47101, CVE-2026-47102, CVE-2026-40217)](https://obsidiansecurity.com/blog/litellm-privilege-escalation-rce)
- [Pentest-Tools: REDCap >= 13.3.0 Unauthenticated Remote Code Execution via Survey Passthru (CVE-2026-90817)](https://pentest-tools.com/vulnerabilities-exploits/redcap-1330-unauthenticated-remote-code-execution-via-survey-passthru_29886)
- [SentinelOne Vulnerability Database: CVE-2026-90817](https://www.sentinelone.com/vulnerability-database/cve-2026-90817/)
- [SentinelOne Vulnerability Database: CVE-2024-10441](https://www.sentinelone.com/vulnerability-database/cve-2024-10441/)
- [GBHackers: Critical N-able N-central Flaw](https://gbhackers.com/critical-n-able-n-central-flaw/)
- [Red Hat Bugzilla: CVE-2026-40217 LiteLLM arbitrary code execution](https://bugzilla.redhat.com/show_bug.cgi?id=2457301)

## Full data

The other 8 detection rules (Splunk SPL / Microsoft KQL / Sigma) unlock with TL Creds; a free account includes daily creds: https://threadlinqs.com/?modal=signup
Indicator values unlock with TL Creds, at a lower price on Red and Purple; AI agents can query rules and IOCs through the Threadlinqs MCP server (Purple): https://intel.threadlinqs.com/mcp

Canonical: https://intel.threadlinqs.com/threat/TL-2026-6450
