Threat reportThreat IntelligenceTL-2026-1704
UT Dallas Study: Multi-Patch CVE Fixes Leave Open Source Exposed to N-Day Exploitation Windows
A University of Texas at Dallas study (Qi, Li, Wang; ESORICS 2026) manually examined 1,646 multi-patch fix records among open-source CVEs in NVD (1999-2025) and found 31.7% take more than a day between the first and last commit, creating an N-day window across unpatched branches, release lines, or forks.
Key facts
| Identifiers | TL-2026-1704, CVE-2012-0038, CVE-2023-4226, CVE-2022-2522 |
|---|---|
| Also known as | UT Dallas Multi-Patch Study, Why Not Fix It Once and for All? |
| Severity | MEDIUM |
| Exploitation status | Public proof-of-concept exploit for a linked CVE |
| EPSS | CVE-2023-4226: 2.4% (83.9th percentile); CVE-2022-2522: 0.5% (46.0th percentile); CVE-2012-0038: 0.3% (32.6th percentile) |
| Affected products | Linux kernel community Linux kernel (XFS); Chamilo Association Chamilo LMS; Vim vim/vim; ImageMagick Studio LLC ImageMagick |
| Weaknesses (CWE) | CWE-190, CWE-434, CWE-122 |
| Tools | CodeBERT, Devign, FIRE, LineVul, ReDebug, ReVeal, UniXcoder |
| MITRE ATT&CK techniques | T1005 Data from Local System, T1041 Exfiltration Over C2 Channel, T1068 Exploitation for Privilege Escalation, T1071 Application Layer Protocol, T1190 Exploit Public-Facing Application, T1203 Exploitation for Client Execution, T1204 User Execution, T1210 Exploitation of Remote Services, T1211 Exploitation for Stealth, T1499 Endpoint Denial of Service, T1505 Server Software Component, T1518 Software Discovery and 4 more in the attack flow |
| Motivation | UNKNOWN |
| Target sectors | technology, softwaredevelopment, education, government administration, criticalinfrastructure |
| Target regions | Global |
| Detection rules | 9 (1 shown free) |
| Sources | 14 cited sources |
Detection coverage
As of 2026-07-26, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-1704 across Splunk SPL, Microsoft KQL and Sigma, covering 28 indicator(s) of compromise. The whole corpus is readable without an account. Every threat comes with one complete detection in Splunk SPL, Microsoft KQL and Sigma, free for everyone; a free account adds daily TL Creds to unlock the rest. Paid plans add a larger daily allowance, correlation and the MCP server. Threadlinqs MCP server · View plans.
Free detection: Unrestricted Upload Paired With .htaccess-Style Override in App Cache (CVE-2023-4226 Pattern) — SPL
Detects the CVE-2023-4226-class (Chamilo LMS work.ajax.php) post-exploitation artifact pattern from research.json: an authenticated low-privilege (learner-role-equivalent) user uploads a script-extension payload into an application cache/upload directory together with a new or modified .htaccess-style override in the same directory, consistent with a workaround-only patch (first commit) that has not yet received the fully-closing follow-up commit ~16 days later. This combination is a strong web-shell-persistence signal regardless of which specific application is affected.
- Languages: Splunk SPL, Microsoft KQL, Sigma
- ATT&CK: T1505.003, T1552.001, T1190
- Data sources: linux_auditd, file_integrity_monitoring, web_server_logs
- Severity: critical
- Confidence: high
Splunk SPL
index=linux_fs sourcetype="linux:auditd:file" (path="*/app/cache/*" OR path="*/uploads/*")
| eval dir=replace(path, "/[^/]+$", "")
| eval file_type=case(
match(path, "(?i)\.(php\d?|phtml|pht|phar)$"), "script_upload",
match(path, "(?i)\.htaccess$"), "htaccess_override",
1=1, "other")
| where file_type!="other"
| transaction dir maxspan=5m startswith=(file_type="script_upload") endswith=(file_type="htaccess_override")
| eval detection="unrestricted_upload_plus_htaccess_override"
| table _time, dir, path, file_type, user, uid, exe, duration
| sort - _timeMicrosoft KQL
DeviceFileEvents
| where Timestamp > ago(24h)
| where FolderPath has "/app/cache/" or FolderPath has "/uploads/" or FolderPath has @"\app\cache\"
| where FileName endswith ".php" or FileName endswith ".phtml" or FileName endswith ".pht" or FileName endswith ".phar" or FileName has_cs ".htaccess"
| summarize UploadEvents = make_set(pack("File", FileName, "Time", Timestamp, "Account", InitiatingProcessAccountName)), ScriptUpload = countif(FileName endswith ".php" or FileName endswith ".phtml" or FileName endswith ".pht" or FileName endswith ".phar"), HtaccessOverride = countif(FileName has_cs ".htaccess") by DeviceName, FolderPath, bin(Timestamp, 5m)
| where ScriptUpload >= 1 and HtaccessOverride >= 1
| project DeviceName, FolderPath, ScriptUpload, HtaccessOverride, UploadEventsSigma
title: Unrestricted Upload Artifact Paired With .htaccess-Style Override in App Cache/Upload Directory
id: tl-2026-1704-linux-webshell-marker
status: experimental
description: Detects a script-extension file (.php/.phtml/.pht/.phar) landing in an application's cache or upload directory alongside a newly created or modified .htaccess-style override file in the same directory within a short window -- consistent with the CVE-2023-4226-class unrestricted-upload-plus-override pattern that a workaround-only first commit fails to fully close.
references:
- https://nvd.nist.gov/vuln/detail/CVE-2023-4226
- https://starlabs.sg/advisories/23/23-4226
author: Threadlinqs Team
date: 2026/07/26
tags:
- attack.initial-access
- attack.t1190
- attack.persistence
- attack.t1505.003
- attack.credential-access
- attack.t1552.001
- cve.2023.4226
logsource:
product: linux
category: file_event
detection:
selection_file:
TargetFilename|contains:
- '/app/cache/'
- '/uploads/'
TargetFilename|endswith:
- '.php'
- '.phtml'
- '.pht'
- '.phar'
selection_htaccess:
TargetFilename|endswith: '.htaccess'
condition: selection_file or selection_htaccess
falsepositives:
- Legitimate deployment/CI pipeline pushing a new .htaccess and cache-warming files to app/cache during a scheduled release window; correlate the two selections and exclude known deployment service accounts
level: highKnown false positives:
- Legitimate deployment/CI pipeline pushing a new .htaccess and cache-warming files to app/cache during a scheduled release window.
Tuning notes: Exclude known CI/CD deployment service accounts and their expected cache-directory write patterns; require the script-extension file and the .htaccess-style file to originate from the same non-deployment user/session and from an authenticated low-privilege role (e.g. learner/student) for higher fidelity.
- Cross-Branch Signed/Unsigned Bounds-Check Bypass → Anomalous Child Process (CVE-2012-0038 Pattern) — KQL · unlock with TL Creds
- Unrestricted Upload Paired With .htaccess-Style Override in App Cache (CVE-2023-4226 Pattern) — KQL · unlock with TL Creds
- Repeated Cross-Branch/Fork Version-Fingerprint Probing With Outbound Report Callback — KQL · unlock with TL Creds
- Cross-Branch Signed/Unsigned Bounds-Check Bypass → Anomalous Child Process (CVE-2012-0038 Pattern) — Sigma · unlock with TL Creds
- Unrestricted Upload Paired With .htaccess-Style Override in App Cache (CVE-2023-4226 Pattern) — Sigma · unlock with TL Creds
- Repeated Cross-Branch/Fork Version-Fingerprint Probing With Outbound Report Callback — Sigma · unlock with TL Creds
- Cross-Branch Signed/Unsigned Bounds-Check Bypass → Anomalous Child Process (CVE-2012-0038 Pattern) — SPL · unlock with TL Creds
- Repeated Cross-Branch/Fork Version-Fingerprint Probing With Outbound Report Callback — SPL · unlock with TL Creds
Indicators of compromise
As of 2026-07-26, Threadlinqs Intelligence tracks 28 indicators of compromise for TL-2026-1704.
Sample indicators (3 of 28)
- URL:
hxxps://arxiv[.]org/abs/2607.13206 - File hash:
093019cf1b18dd31b2c3b77acce4e000e2cbc9ce - URL:
hxxps://starlabs[.]sg/advisories/23/23-4226
Defanged, TLP:CLEAR. Full set of 28 indicators: unlock with TL Creds.
Technical analysis
Seven ML-based incomplete-fix detectors (CodeBERT, UniXcoder, LineVul, Devign, ReVeal) scored under 50% accuracy/F1, and clone detectors ReDebug and FIRE degraded sharply across branches (FIRE's true-positive rate fell from ~90% to ~52%), meaning defenders cannot rely on automated tooling to flag an incomplete patch.
Researchers Weiliang Qi, Youpeng Li, and Xinda Wang at the University of Texas at Dallas published "Why Not Fix It Once and for All? An Empirical Study of Multiple Patches for Vulnerability Fixes in Open-Source Software" (arXiv:2607.13206, accepted ESORICS 2026), manually reviewing 1,646 open-source CVE records from NVD (1999-2025) whose fixes spanned more than one commit -- about 6.7% of all open-source CVEs with linked patches, averaging 2.55 patches per CVE.
The study builds a three-category, six-subcategory taxonomy for why fixes fragment: (1) the same defect exists in multiple code locations, branches, or downstream forks requiring distributed patches -- ImageMagick alone accounted for roughly one-third of its fixes this way due to parallel release-line maintenance; (2) ancillary, non-security commits (documentation edits, version bumps, temporary workarounds) get bundled with or precede the real fix, affecting 5.8% of cases; and (3) the initial patch was incomplete or defective and reintroduced or failed to close the flaw -- 641 of 1,646 records, the second-largest subcategory.
Critically, 31.7% of multi-patch fixes took more than 24 hours between the first and final corrective commit. During that window, a vulnerability disclosed or partially patched on one branch remains exploitable on sibling branches, release lines, or downstream forks that have not yet received the follow-up commit -- an N-day exposure that exists purely because of patch fragmentation, independent of any vendor delay in initial disclosure.
Three real CVEs illustrate each pattern. CVE-2012-0038 (Linux kernel, CWE-190, integer overflow in xfs_acl_from_disk): the first commit (093019cf1b18dd31b2c3b77acce4e000e2cbc9ce) added a bounds check but used a helper returning an unsigned int against a signed count variable, letting a negative count bypass the check; a second, corrective commit (fa8b18edd752a8b4e9d1ee2cd615b82c93cf8bba) landed roughly 18 days later, converting the variable to unsigned and closing the bypass -- an 18-day window during which the initial "fixed" kernel remained vulnerable to a local DoS/heap corruption via a malformed on-disk ACL. CVE-2023-4226 (Chamilo LMS, CWE-434, unrestricted file upload in /main/inc/ajax/work.ajax.php): STAR Labs SG researcher Ngo Wei Lin disclosed to the vendor on 2023-09-04 that an authenticated learner-role user could upload a PHP web shell plus a crafted .htaccess into the app cache directory to obtain RCE; Chamilo shipped a workaround-level commit before a complete closing fix followed roughly 16 days later, with the full release (v1.11.26) and public disclosure landing 2023-09-27 and 2023-11-28 respectively. CVE-2022-2522 (Vim, CWE-122, heap-based buffer overflow, published 2022-07-25): the study found the accepted fix retained an unrelated, non-security commit alongside the real change, and after a user flagged the residual defect, the vendor never issued a further correction -- the case sits permanently in the "incomplete/never corrected" bucket rather than resolving within days.
None of the three example CVEs appear in the CISA Known Exploited Vulnerabilities catalog as of 2026-07-26; they are presented as illustrative, patched (or permanently unresolved) case studies of the fragmentation phenomenon rather than active in-the-wild campaigns. The operational risk is structural and forward-looking: any future multi-branch CVE has roughly a 1-in-3 chance of leaving a same-day-undetectable N-day gap that current commercial and open-source incomplete-fix detectors are not equipped to flag.
MITRE ATT&CK techniques by tactic:
Collection
Exfiltration
T1041 Exfiltration Over C2 Channel
Privilege Escalation
T1068 Exploitation for Privilege Escalation
Command and Control
T1071 Application Layer Protocol
Initial Access
T1190 Exploit Public-Facing Application
Execution
T1203 Exploitation for Client Execution; T1204 User Execution
Lateral Movement
T1210 Exploitation of Remote Services
Defense Evasion
T1211 Exploitation for Stealth
Impact
T1499 Endpoint Denial of Service
Persistence
T1505 Server Software Component
Discovery
Credential Access
Resource Development
Reconnaissance
Remediation
Patches
- Linux kernel >= 3.1.9 (CVE-2012-0038)
- Chamilo LMS >= 1.11.26 (CVE-2023-4226)
- Vim >= 9.0.0061 (CVE-2022-2522)
Immediate actions
- Treat the first published commit/patch for any multi-branch or multi-fork open-source CVE as provisional -- confirm whether NVD/vendor references list more than one corrective commit before marking the CVE closed
- For CVE-2012-0038-class Linux kernel deployments, confirm kernel >= 3.1.9 (or backport of commit fa8b18edd752a8b4e9d1ee2cd615b82c93cf8bba, not just 093019cf1b18dd31b2c3b77acce4e000e2cbc9ce)
- For Chamilo LMS, confirm version >= 1.11.26 (CVE-2023-4226 fully closed); versions <= 1.11.24 with only a partial workaround remain exploitable via /main/inc/ajax/work.ajax.php
- For Vim, confirm version >= 9.0.0061 for CVE-2022-2522; note the study found the shipped fix still carries an unrelated commit, so re-verify against upstream if replicating this analysis
Workarounds
- Restrict learner-role upload paths and disable .htaccess override in Chamilo's app/cache directory as a compensating control if unable to patch immediately
- Disable untrusted ACL-bearing filesystem mounts on unpatched Linux kernels vulnerable to CVE-2012-0038
Longer-term hardening
- Do not rely solely on CodeBERT/UniXcoder/LineVul/Devign/ReVeal-class ML incomplete-fix classifiers or ReDebug/FIRE-class clone detectors as a sole gate for patch-completeness verification -- all scored below 50% accuracy/F1 or degraded sharply (FIRE: ~90% to ~52% TPR) across branches in the study
- Build patch-verification workflow steps that explicitly check for follow-up commits referencing the original CVE/commit hash across all maintained branches and downstream forks before closing a vulnerability-management ticket
- Track average patch latency for multi-commit CVEs in your dependency tree; prioritize re-scanning branches/release lines that historically lag on follow-up commits (e.g., projects like ImageMagick with heavy parallel release-line maintenance)
Affected and fixed versions
- Linux kernel community — Linux kernel (XFS)
Vulnerable versions: before 3.1.9
Fixed in: 3.1.9 and later (commit fa8b18edd752a8b4e9d1ee2cd615b82c93cf8bba) - Chamilo Association — Chamilo LMS
Vulnerable versions: up to 1.11.24
Fixed in: 1.11.26 and later - Vim — vim/vim
Vulnerable versions: before 9.0.0061
Fixed in: 9.0.0061 and later - ImageMagick Studio LLC — ImageMagick
Vulnerable versions: study finding is aggregate, not tied to a single CVE/version: ~30% of ImageMagick security fixes required multiple commits across parallel release lines
Fixed in: varies per release line; not enumerated by the study
Timeline
- CVE-2012-0038 discussed on the oss-security mailing list; initial Linux kernel commit locked-ioc-2 adds a bounds check to xfs_acl_from_disk but uses a helper returning an unsigned int against a signed count, leaving the check bypassable.
- Corrective follow-up commit locked-ioc-3 lands roughly 18 days after the first, converting the ACL count variable to unsigned and closing the integer-overflow bypass in xfs_acl_from_disk.
- CVE-2012-0038 formally published in the National Vulnerability Database.
- CVE-2022-2522 (Vim heap-based buffer overflow, CWE-122) published; the UT Dallas study found the accepted fix retained an unrelated, non-security commit and the flaw was never further corrected after a user flagged it to the vendor.
- STAR Labs SG researcher Ngo Wei Lin discloses the Chamilo LMS work.ajax.php unrestricted file upload (later CVE-2023-4226) to the vendor.
- Chamilo LMS v1.11.26 released with the complete closing fix; the UT Dallas study identifies this as the second corrective commit, landing roughly 16 days after an initial workaround-only patch.
- CVE-2023-4226 publicly disclosed and published in NVD and GitHub Security Advisories.
- Qi, Li, and Wang submit "Why Not Fix It Once and for All?" (arXiv:2607.13206) to arXiv, accepted at ESORICS 2026, documenting the 1,646-record multi-patch CVE study.
- Help Net Security publishes coverage of the UT Dallas multi-patch study, surfacing the N-day exposure-window finding and detection-tooling gap to SOC and patch-management audiences.
Sources cited
- NVD - CVE-2012-0038 NVD (NIST) · Advisory · 2012-03-16
- NVD - CVE-2023-4226 NVD (NIST) · Advisory · 2023-11-28
- NVD - CVE-2022-2522 NVD (NIST) · Advisory · 2022-07-25
- Why Not Fix It Once and for All? An Empirical Study of Multiple Patches for Vulnerability Fixes in Open-Source Software arxiv.org · Research · 2026-07-14
- Red Hat Bugzilla #773280 (xfs_acl_from_disk integer overflow) bugzilla.redhat.com · Vendor report · 2012-01-10
- (CVE-2023-4226) Chamilo LMS Work Ajax File Upload Functionality Remote Code Execution starlabs.sg · Proof of concept · 2023-11-28
- Linux kernel commit 093019cf1b18dd31b2c3b77acce4e000e2cbc9ce (initial xfs_acl_from_disk fix) GitHub · Code repository · 2012-01-10
- Linux kernel commit fa8b18edd752a8b4e9d1ee2cd615b82c93cf8bba (corrective follow-up fix) GitHub · Code repository · 2012-01-28
- Vim patch commit 5fa9f23a63651a8abdb074b4fc2ec9b1adc6b089 GitHub · Code repository · 2022-07-25
- Vim patch commit b9e717367c395490149495cf375911b5d9de889e GitHub · Code repository · 2022-07-25
- Huntr bounty report for Vim heap-based buffer overflow (CVE-2022-2522) huntr.dev · Proof of concept · 2022-07-25
- kqueue.org: CVE-2012-0038 XFS ACL count integer overflow kqueue.org · Blog · 2012-01-10
- Multi-patch vulnerability fixes can leave open source exposed Help Net Security · News · 2026-07-23
- Multiple Patch Fixes Pose Security Risks to Open Source Projects news4hackers.com · News · 2026-07-23
Previous and next in threat intel
Drafted by the Threadlinqs research pipeline from 14 sources; published ; last updated —.