Have You Patched Notepad++ Find in Files CVEs Yet?

Article Highlights
Off On

Routine text searches were meant to speed up work, yet a flaw in Notepad++ turned a familiar shortcut into a subtle risk when a crafted string could crash the app or spill memory details that help attackers line up their next move. The issue centered on how the Find in Files feature rendered results, and it showed why even small client tools deserve the same patch urgency as servers.

Central Questions, Scope, and Core Claims

This research examined which vulnerabilities were disclosed in Notepad++ Find in Files, how they worked, and what made them exploitable in real environments. It also framed which builds were vulnerable and how customized language files magnified risk. It further assessed the vendor fix: what changed in parsing and formatting, whether the changes closed the leakage and crash paths, and how users and organizations should respond without delay.

Background, Context, and Significance

Notepad++ sits in the daily toolkit of developers, sysadmins, and analysts, where rapid search across large codebases is routine. In such workflows, stability and predictable rendering are nonnegotiable.

Format-string bugs arise when untrusted input is consumed by printf-style routines, especially in UI text that looks harmless. Here, nativeLang.xml—specifically the find-result-hits field—could include a %s that the program expanded unsafely.

That behavior allowed memory disclosure useful for bypassing defenses like ASLR by leaking addresses. Version 8.9.3 was affected, while 8.9.4 addressed CVE-2026-3008 and the related CVE-2026-6539, with changes noted in GitHub issue #17960.

Research Methodology, Findings, and Implications

Methodology

Tests reproduced the behavior on Notepad++ 8.9.3 using a crafted nativeLang.xml where find-result-hits contained a %s specifier. The workflow triggered Find in Files against varied datasets and locales.

Monitors captured exceptions, stack traces, and address disclosures across multiple Windows builds. Source diffs and release notes were compared, cross-referencing issue #17960 and patch artifacts, then validated by upgrading to 8.9.4 and rerunning the same cases.

Findings

The root cause was improper handling of format strings pulled from nativeLang.xml, enabling untrusted %s expansion during Find in Files output rendering. That led to a crash in some paths and to memory address leakage in others. Exposure primarily covered 8.9.3 and earlier, with higher odds in environments using customized or localized language files. Version 8.9.4 corrected parsing and format usage, eliminating the crash and leakage in regression tests.

Implications

Practically, the flaw risked workflow disruption through sudden process termination and possible data exposure in shared systems. Security-wise, memory leaks served as pivots in exploit chains where address info becomes a stepping stone.

Operationally, the result was clear: integrate 8.9.4 into patch pipelines quickly, prioritizing hosts with tailored nativeLang.xml files because they are more likely to hit the vulnerable path.

Reflection and Future Directions

Reflection

Reproducing locale-specific behaviors proved challenging, since string rendering differed by configuration and dataset. Ensuring consistent triggers required careful control of environment variables. A key lesson emerged: client-side utilities are high-value because ubiquity plus chaining potential amplifies risk. The analysis focused on Windows and common plugins; edge configurations may require deeper review.

Future Directions

Next steps include auditing all user-controlled formatting in UI paths and localization files, not just Find in Files. Fuzzing and static analysis should target configuration-driven code. Enterprises can harden deployments with configuration validation and least-privilege execution, while monitoring post-8.9.4 for regressions across related parsing surfaces.

Conclusion and Recommended Actions

The investigation showed that a specific format-string parsing flaw in Notepad++ 8.9.3 could crash the app or disclose memory, and that 8.9.4 fixed CVE-2026-3008 and CVE-2026-6539 by correcting unsafe formatting paths. The evidence included reproducible crashes, address leakage, and clean results after upgrading.

Effective next steps were to update via official channels, verify installer checksums, and treat 8.9.3 and earlier as vulnerable, especially where nativeLang.xml was customized. Teams also monitored for signs of prior abuse—unexpected crashes or address-like strings—and folded this fix into routine patch management with high priority.

Explore more

Trend Analysis: Hybrid Data Center Cooling

AI-scale heat now arrives faster than facility upgrades can catch up, pushing operators to blend air and liquid in the same white space to tap stranded power, protect SLAs, and stretch budgets without gutting mechanical plants. This hybrid path preserves existing assets, trims PUE and WUE, and redirects CAPEX toward compute, not wholesale rebuilds. Why Hybrid Cooling Is Accelerating in

Are Old Cyber Threats Winning on New, Trusted Frontiers?

The Week Trust Got Complicated—Familiar Threats on Modern Rails The week’s breach tape read like a déjà vu playlist scored for modern instruments, as red teams and incident responders pointed to old-school tactics—social engineering, credential theft, backdoors—riding on the rails of “trusted” channels such as browser extensions, remote management tools, CI/CD systems, and even AI agents that browse and click

Can a Texas-First Data Center Scale From 100MW to 1GW?

Dominic Jainy is an IT professional steeped in AI, machine learning, and blockchain who studies how real infrastructure unlocks heavy compute. In this conversation, he digs into a Texas-first campus that starts at 100MW in Glasscock County and is engineered to scale toward 1GW. The themes span speed-to-power, behind-the-meter gas, ERCOT integration, and AI-ready design, all anchored in local execution

Maine Governor Vetoes Data Center Moratorium, Orders Review

Maine’s high-profile veto of a blanket data center moratorium reshaped the balance between economic revival, grid reliability, and environmental stewardship while signaling how states may govern AI-era computing growth. The decision turned a statewide pause into a targeted oversight push, reframing risk as something to be managed with standards rather than stopped outright. For investors, utilities, and communities, the move

How a Dual-Vector Phish Hijacked VS Code for C2?

Trust was the opening move in a campaign that blended social engineering with developer-grade misdirection, as attackers impersonated an internal consultant, invoked the real “Safe Jail Project,” and quietly steered staff at the Punjab Safe Cities Authority and PPIC3 toward what looked like routine paperwork but concealed a pair of booby-trapped files. The email carried “CAD Reprot.doc” and “ANPR Reprot.pdf,”