CVE-2026-106620
Executive Summary
CVE-2026-106620 is a high-severity Incorrect Authorization vulnerability (CWE-863) in ColdFusion that enables a low-privileged attacker to bypass security controls and obtain unauthorized write access. With a CVSS v3.1 base score of 9.6 (AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N), the flaw is remotely exploitable, requires only low privileges, needs no user interaction, and — critically — changes scope, meaning the compromised component can affect resources beyond its original security authority. Adobe ColdFusion is widely deployed for public-facing web applications and internal business automation, and no patch, workaround, public proof-of-concept, or CISA KEV entry currently exists for this CVE, leaving defenders reliant on compensating controls and detection engineering.
This record — and 170,000+ others — are available via the ValtersIT API. One endpoint replaces five: NVD + GitHub CVE + CISA KEV + EPSS + Nuclei + Exploit-DB, unified.
No signup required
Technical Analysis
The vulnerability is classified as CWE-863: Incorrect Authorization. This weakness class covers cases where an application performs an authorization check but does so incorrectly — for example, validating the existence of a session or a role but not validating that the authenticated principal is actually entitled to perform the requested write operation, or applying authorization logic to one code path while an alternative path reaches the same sensitive operation unprotected.
The NVD description states that ColdFusion is affected by an incorrect authorization flaw that "could result in a Security feature bypass" and that a "low-privileged attacker could leverage this vulnerability to bypass security measures and gain unauthorized write access." Exploitation "does not require user interaction."
From these statements, the following is derivable with confidence:
- Attack surface: ColdFusion is a web application server; the vulnerability is network-reachable (AV:N) and the affected interface is presumed to be a request-handling endpoint of the ColdFusion application. The specific servlet, template, endpoint, or API path is not identified in the available data.
- Privilege requirement: PR:L indicates the attacker must possess some valid low-privileged foothold (an authenticated session, a low-privilege role, or an equivalent credential). This is not a pre-authentication flaw as scored.
- Mechanism: The authorization decision protecting a write operation is incorrectly enforced, allowing a low-privileged principal to reach functionality reserved for a higher-privileged principal. The precise misconfiguration — for example role-check omission, path normalization mismatch in an authorization filter, or an insecure direct object reference combined with a weak ownership check — cannot be determined from the available description and is not asserted here.
- Scope change (S:C): The successful bypass impacts resources outside the vulnerable component's security scope. In a ColdFusion context this may mean the attacker reaches the underlying host filesystem, the application server's administrative authority, or downstream systems the ColdFusion instance trusts — but the specific cross-boundary resource is not specified in the source data.
- Confidentiality/Integrity impact: C:H and I:H with A:N. The description explicitly cites "unauthorized write access," consistent with the high integrity impact. High confidentiality impact is scored, implying the same bypass class also exposes protected data; the description does not enumerate what data.
No function names, class names, endpoint paths, or parameter names are provided in the CNA title or NVD description, so none are asserted in this report. The CNA title field is empty. Affected version ranges are not published by any tracked source, and no OSV package data is available. Until Adobe publishes an advisory with version ranges, assume all supported ColdFusion releases in production are candidates pending vendor confirmation.
Because this is an authorization-logic defect rather than a memory-safety bug, exploitation typically manifests as a well-formed request that simply should have been rejected. That characteristic makes network detection difficult: the malicious request often looks structurally identical to legitimate traffic and differs only in who is sending it and what they are allowed to change.
Exploitation Scenarios
No public proof-of-concept, Nuclei template, or exploit code exists for CVE-2026-106620. The attack chain below is therefore described at the vulnerability-class level, inferred strictly from CWE-863 and the NVD description. Specific endpoint names, parameters, and payload syntax are intentionally omitted because they are not present in the source data.
1) Reconnaissance and target identification. The attacker identifies internet- or intranet-exposed ColdFusion instances. Fingerprinting relies on ColdFusion-specific response artifacts, default administrative paths, cookie and header patterns, and version disclosure. An attacker with an existing low-privileged account (or the ability to self-register or obtain one through legitimate or social means) maps the application's write-capable operations — content creation, configuration changes, scheduled tasks, or data submission — and observes which of them return authorization errors.
2) Triggering the authorization defect. With a low-privileged session in hand, the attacker issues requests to a write-capable operation that the application should restrict to a higher-privileged role. The defect causes the authorization check on that operation to evaluate incorrectly — the check may be absent on the reachable code path, may validate authentication but not authorization, or may be bypassed because the request is normalized or routed in a way the check does not cover. The key signature of CWE-863 is that the request is syntactically acceptable and the server processes it rather than rejecting it.
3) Execution method at the vulnerability-class level. No memory corruption or code injection is implied. The "exploitation" is disciplined use of a legitimate-looking, correctly formed request that exceeds the caller's authority. Payload construction requires only knowledge of the target operation and the attacker's own low-privileged session material. No user interaction is required on the victim side, consistent with UI:N.
4) Post-exploitation impact. Successful exploitation yields unauthorized write access. Because the CVSS vector scores scope as changed, the writes can affect assets outside the immediate vulnerable component — for example, modifying server-side content or configuration that alters behavior of the ColdFusion server itself or of systems that consume its output. The specific destructive, persistence, or lateral-movement actions available are not described in the source data and are not speculated here. The high confidentiality score also implies the same bypass technique may expose protected data.
Impact Assessment
- Base score: 9.6 (Critical), CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N.
- Exploitability subscores: AV:N and AC:L mean the flaw is exploitable across a network with no special conditions; PR:L lowers the bar to any party holding a low-privileged account; UI:N means no victim action is needed, enabling automated, unattended exploitation once a session exists.
- Impact subscores: C:H and I:H reflect severe confidentiality and integrity loss; A:N indicates availability is not directly impacted by the scored vector.
- Scope: S:C is the most operationally significant element. A scope-changing bypass means containment cannot be assumed at the ColdFusion application boundary — defenders must treat the affected host and anything it trusts as potentially in-blast-radius.
- Business impact: ColdFusion frequently fronts customer portals, internal workflow systems, and integrations. Unauthorized write access can mean tampered content, altered business logic, corrupted records, or poisoned downstream data. Because no patch exists, mean time to remediate is unbounded, and exposure persists until Adobe ships a fix and organizations deploy it.
- Exploitation likelihood: EPSS has not yet scored this CVE, and it is not in CISA KEV. No public exploit exists. Current real-world exploitation risk is lower than the raw 9.6 score would suggest, but historical ColdFusion CVE exploitation (rapid weaponization of deserialization and authorization flaws) means the risk can shift quickly once technical details or a PoC surface.
Detection Guidance
The detection content below targets the behavioral signature of an authorization bypass: successful, state-changing requests originating from low-privileged sessions against ColdFusion administrative or write-capable endpoints. Rule logic is generic by necessity, because the vulnerable endpoint is not published. Tune path lists and role/session field names to your environment.
Sigma Rule
title: Potential ColdFusion Incorrect Authorization Write Bypass (CVE-2026-106620)
id: 7c9e2a54-1f6b-4d3e-9a01-coldfusion106620
status: experimental
description: Detects successful state-changing requests to ColdFusion administrative or write-capable endpoints from low-privileged sessions, consistent with CWE-863 authorization bypass.
references:
- https://nvd.nist.gov/vuln/detail/CVE-2026-106620
author: valtersit.com Threat Intelligence
date: 2026/01/01
logsource:
category: webserver
detection:
selection_method:
cs-method:
- POST
- PUT
- PATCH
selection_cf_paths:
cs-uri-stem|contains:
- '/CFIDE/'
- '/cfide/'
- '/administrator/'
- '/adminapi/'
selection_success:
sc-status:
- 200
- 201
- 204
selection_lowpriv_indicator:
cs-uri-query|contains:
- 'role='
- 'admin='
- 'isAdmin='
filter_trusted_admin_networks:
c-ip|startswith:
- '10.'
- '192.168.'
- '172.16.'
condition: selection_method and selection_cf_paths and selection_success and selection_lowpriv_indicator and not filter_trusted_admin_networks
falsepositives:
- Legitimate administrative activity from non-listed network ranges
- Authenticated power users performing authorized writes
level: high
tags:
- attack.t1190
- cve.2026-106620
YARA Rule
rule CVE_2026_106620_ColdFusion_Authz_Bypass
{
meta:
description = "Detects ColdFusion application/log artifacts indicating attempted authorization bypass on write operations (CVE-2026-106620)"
author = "valtersit.com Threat Intelligence"
reference = "https://nvd.nist.gov/vuln/detail/CVE-2026-106620"
date = "2026-01-01"
severity = "high"
strings:
$cfide = "/CFIDE/" nocase
$admin = "/administrator/" nocase
$authz1 = "authorization" nocase
$authz2 = "access denied" nocase
$bypass = "security feature bypass" nocase
$write = "write access" nocase
condition:
uint16(0) == 0x3C3F or
(
(any of ($cfide, $admin)) and
2 of ($authz1, $authz2, $bypass, $write)
)
}
Suricata IDS Rule
alert http any any -> any any (msg:"CVE-2026-106620 ColdFusion Incorrect Authorization Write Bypass Attempt"; flow:established,to_server; http.method; content:"POST"; http.uri; content:"/CFIDE/"; nocase; http.uri; content:"/administrator/"; nocase; distance:0; http.header; content:"Cookie"; nocase; classtype:attempted-admin; reference:cve,2026-106620; reference:url,nvd.nist.gov/vuln/detail/CVE-2026-106620; metadata:created_at 2026_01_01, updated_at 2026_01_01; sid:2026106620; rev:1;)
alert http any any -> any any (msg:"CVE-2026-106620 ColdFusion Authz Bypass - Write Method To Admin Path"; flow:established,to_server; http.method; content:"PUT"; http.uri; content:"/CFIDE/"; nocase; classtype:attempted-admin; reference:cve,2026-106620; sid:2026106621; rev:1;)
Elastic Detection
sequence by source.ip with maxspan=5m
[ network where http.request.method in ("POST", "PUT", "PATCH") and
url.path : ("/CFIDE/*" or "/administrator/*") and
network.protocol == "http" ]
[ network where http.response.status_code in (200, 201, 204) and
url.path : ("/CFIDE/*" or "/administrator/*") and
not source.ip : ("10.0.0.0/8" or "172.16.0.0/12" or "192.168.0.0/16") ]
Splunk SPL Query
index=web sourcetype=access_combined
| eval is_cf_path=if(match(uri_path, "(?i)/(CFIDE|cfide|administrator|adminapi)/"), 1, 0)
| where is_cf_path=1 AND method IN ("POST","PUT","PATCH") AND status IN (200,201,204)
| eval is_lowpriv=if(match(uri_query, "(?i)(role|admin|isAdmin)="), 1, 0)
| where is_lowpriv=1
| stats count min(_time) as first_seen max(_time) as last_seen values(uri_path) as paths values(status) as statuses by src_ip
| where count > 3
| convert ctime(first_seen) ctime(last_seen)
| sort - count
Wazuh OSSEC Rule
<group name="web,coldfusion,cve-2026-106620,">
<rule id="106620" level="12">
<decoded_as>web-accesslog</decoded_as>
<field name="url">/CFIDE/|/cfide/|/administrator/|/adminapi/</field>
<regex>"(POST|PUT|PATCH) .*(CFIDE|administrator|adminapi)"</regex>
<description>CVE-2026-106620: ColdFusion state-changing request to admin/write endpoint - possible incorrect authorization bypass</description>
<mitre>
<id>T1190</id>
</mitre>
<group>attack,coldfusion_authz_bypass,</group>
</rule>
<rule id="106621" level="14">
<if_sid>106620</if_sid>
<regex>" (200|201|204) </regex>
<description>CVE-2026-106620: ColdFusion admin/write endpoint returned success status - probable authorization bypass success</description>
<mitre>
<id>T1190</id>
</mitre>
<group>attack,coldfusion_authz_bypass,</group>
</rule>
</group>
Mitigation & Patch Guidance
Patch confidence: 0/4 — UNPATCHED. No vendor advisory, patch reference, fixed version, or workaround is published by any tracked source as of this report. Affected version ranges are unknown, and no OSV package data exists. The steps below are compensating controls and hardening measures, not substitutes for a vendor fix.
1) Immediate mitigations
- Restrict network access to ColdFusion administrative and write-capable interfaces. Where possible, allow only trusted management networks or VPN egress; do not expose the ColdFusion administrative interface to the internet.
- Enforce authentication and authorization at a reverse proxy or WAF in front of ColdFusion, applying explicit allow-lists of permitted write operations per role, rather than relying solely on in-application authorization (which is the layer this CVE defeats).
- Audit existing low-privileged accounts. Disable or tightly scope any accounts that do not strictly require write capability, and rotate credentials for administrative and low-privilege roles alike.
- Enable verbose request logging (source IP, authenticated principal, method, path, status) for all write operations and forward it to the SIEM so the detection content above has telemetry to act on.
- Monitor for unexpected writes: content changes, scheduled task creation, configuration modification, and filesystem writes by the ColdFusion service account.
2) Patch / upgrade steps
- Monitor Adobe's security bulletin channel for the ColdFusion advisory addressing CVE-2026-106620. No advisory or fixed version is available at the time of writing.
- When a fixed release ships, test in a staging environment and deploy to all internet- and intranet-exposed ColdFusion instances on an emergency timeline given the CVSS 9.6 rating and absence of a workaround.
- Inventory all ColdFusion deployments, including forgotten, legacy, and developer instances, so the emergency patch can be applied exhaustively. Unmanaged instances are the most likely to be missed and the most likely to be targeted.
3) Long-term hardening
- Adopt deny-by-default authorization: every write operation must fail closed unless the caller's role is explicitly granted that specific operation. Periodically test authorization boundaries with role-matrix regression tests.
- Apply the principle of least privilege to all ColdFusion service accounts and application roles; eliminate shared or generic administrative credentials.
- Establish continuous vulnerability monitoring for ColdFusion (Adobe PSIRT bulletins plus this CVE's NVD record) and subscribe to CISA KEV alerts, since ColdFusion flaws have historically been added to KEV after weaponization.
- Segment ColdFusion servers so that a scope-changing compromise cannot reach downstream trust relationships, backups, or shared storage.
- Conduct regular red-team or purple-team exercises against the role matrix to validate that low-privileged principals cannot reach write-capable operations.
References
- NVD — CVE-2026-106620: https://nvd.nist.gov/vuln/detail/CVE-2026-106620
- CWE-863: Incorrect Authorization — https://cwe.mitre.org/data/definitions/863.html
- MITRE ATT&CK T1190 — Exploit Public-Facing Application: https://attack.mitre.org/techniques/T1190/
- CISA Known Exploited Vulnerabilities Catalog (CVE-2026-106620 not listed at time of writing): https://www.cisa.gov/known-exploited-vulnerabilities-catalog
- Adobe Product Security Incident Response Team (PSIRT) security bulletins: https://helpx.adobe.com/security.html
No threat actor campaign, exploit-in-the-wild evidence, public proof-of-concept, Nuclei template, or exploit code has been identified for CVE-2026-106620. The CVE is not present in CISA KEV, and EPSS has not yet assigned a score.
What is this? Security researchers use these exact search queries on Shodan.io (a search engine for internet-connected devices) to find vulnerable systems globally. You can copy these dorks to see if your organization's IP space is exposed.