CVE-2026-25750Disclosure(langchain / langsmith)

LOWCVSS 8.1 · HIGH

Signal is active with 1 mentions in latest observed window

Immediate actions

Patch langchain langsmith systems immediately

Recommended action window: Monitor and triage in normal cycle

NVD description

Langchain Helm Charts are Helm charts for deploying Langchain applications on Kubernetes. Prior to langchain-ai/helm version 0.12.71, a URL parameter injection vulnerability existed in LangSmith Studio that could allow unauthorized access to user accounts through stolen authentication tokens. The vulnerability affected both LangSmith Cloud and self-hosted deployments. Authenticated LangSmith users who clicked on a specially crafted malicious link would have their bearer token, user ID, and workspace ID transmitted to an attacker-controlled server. With this stolen token, an attacker could impersonate the victim and access any LangSmith resources or perform any actions the user was authorized to perform within their workspace. The attack required social engineering (phishing, malicious links in emails or chat applications) to convince users to click the crafted URL. The stolen tokens expired after 5 minutes, though repeated attacks against the same user were possible if they could be convinced to click malicious links multiple times. The fix in version 0.12.71 implements validation requiring user-defined allowed origins for the baseUrl parameter, preventing tokens from being sent to unauthorized servers. No known workarounds are available. Self-hosted customers must upgrade to the patched version.

langchain CVEs Back to dashboard

0.5/ 10 priority

Sources & remediation

Vendor / third-party advisories

https://github.com/langchain-ai/helm/security/advisories/GHSA-r8wq-jwgw-p74g

Weakness type (CWE)

CWE-74

Priority

LOW

Exploitation

NONE

PoC

NONE

Patch

AVAILABLE

Momentum

STABLE

Are you affected?

If you run products in this scope, you should treat this CVE as relevant to your environment.

langsmith

Threat summary

Patch or workaround signal is available
11 mentions across 7 observed days
Momentum state: stable

What's happening

Patch or workaround mentioned in 5 signals
Technical details provided in 8 signals
Disclosure: 6 classified signals
Peaked 6d ago at 2 mentions (2026-03-04); latest day: 1
11 total mentions across 7 days

Affected systems

Vendors

langchain

Products

langsmith

Deep dive

Activity timeline11 mentions / 7d

Signal classification2 categories

Disclosure

654.5%

Patch

545.5%

Referenced assets11 URLs

By indicator

Classification over time

Date	Total	Labels

2026-03-04	2	Disclosure1Patch1
2026-03-13	2	Disclosure2
2026-03-14	2	Disclosure2
2026-03-15	1	Patch1
2026-03-16	2	Disclosure1Patch1
2026-03-20	1	Patch1
2026-03-29	1	Patch1

Full discourse11 posts

Sekurak@Sekurak195d
Patch
🚨 Wystarczyła manipulacja jednym parametrem żądania, aby przejąć konto w LangSmith. 🛡️ Jeżeli korzystasz z LangSmith do budowy swoich aplikacji AI mamy ważne ostrzeżenie. 💻 Badacze z Miggo Security wykryli poważną lukę bezpieczeństwa (CVE-2026-25750) pozwalającą na przejęcie konta. 🛠️ Źródłem całego zamieszania był parametr baseURL, pozwalający na definicję własnych serwerów backend. Przez brak walidacji danych, przeglądarka użytkownika przesyłała aktywny token sesyjny bezpośrednio na serwer kontrolowany przez atakującego. ✅ Błąd został naprawiony w wersji 0.12.71 LangSmith. Użytkownicy korzystający z architektury chmurowej nie muszą podejmować żadnych działań, poprawka została automatycznie wdrożona. W przypadku rozwiązań self-hosted zalecamy niezwłoczną aktualizację. 👉 Szczegóły: https://sekurak.pl/jak-manipulacja-parametrem-zadania-mogla-doprowadzic-do-przejecia-konta-w-langsmith-szczegoly-podatnosci-cve-2026-25750/
Post summary
CVE‑2026‑25750 in LangSmith permits session hijacking via an unsanitized baseURL parameter; the issue was fixed in version 0.12.71, and users—especially self‑hosted deployments—are urged to update immediately.
0303175.0K
42.8K followersView on X
Audrey Renée Bentley@BentleyAudrey208d
Disclosure
https://api.cyfluencer.com/s/hack-the-ai-brain-uncovering-an-account-takeover-vulnerability-in-langsmith-25937/1 Miggo Security discovered a critical account takeover vulnerability (CVE-2026-25750) in LangSmith. Learn how this flaw exposed proprietary AI data.
Post summary
The excerpt announces the discovery of a critical account takeover flaw (CVE‑2026‑25750) in LangSmith, highlighting its potential to expose proprietary AI data.
0501211.1K
33.2K followersView on X
iototsecnews@iototsecnews204d
Patch
LangSmith の脆弱性 CVE-2026-25750 が FIX：API 設定不備とセッション情報の流出 https://iototsecnews.jp/2026/03/14/critical-langsmith-account-takeover-vulnerability-puts-users-at-risk/ LangSmith で見つかった CVE-2026-25750 は、データの取得先を開発者が自由に指定できる、base URL パラメータの仕様に起因します。この機能において、入力された接続先ドメインが正しいものどうかを確認する検証プロセスが欠けていたことで、アプリケーションが外部からの入力値を無条件に信頼してしまうという問題が生じています。この検証不足により、認証済みのユーザーが細工されたリンクを踏むだけで、ブラウザがセッション認証情報を攻撃者のサーバへ自動送信してしまうリスクが生じています。現在は、許可されたオリジン (信頼できるドメイン) のみを事前登録して制限する対策が取られています。ご利用のチームは、ご注意ください。 #AI #ML #CVE202625750 #LangSmith #Vulnerability
Post summary
The post explains CVE-2026-25750 in LangSmith, detailing the flaw and the vendor’s fix that enforces a whitelist of trusted origins, but it offers no PoC, exploit code, or evidence of active exploitation.
01000173
484 followersView on X
SecAlerts@SecAlertsCo208d
Patch
CVSS 8.5 vuln in LangSmith could enable token theft and full account takeover. Extensive info, including fix info, at SecAlerts: CVE-2026-25750, CVSS 8.5: https://secalerts.co/vulnerability/CVE-2026-25750 #ciso #cio #cto #vulnerabilities #cybersecurity #msp #mssp #secalerts #CVE202625750 #LangSmith https://t.co/6UioUco80I
Post summary
The tweet announces a high‑severity vulnerability in LangSmith (CVE‑2026‑25750) that could allow token theft and account takeover, and directs readers to SecAlerts for fix information.
00010135
803 followersView on X
Ben Rothke@benrothke211d
Disclosure
Critical account takeover vulnerability (CVE-2026-25750) in #LangSmith platform for LLM applications discovered by @MiggoSecurity. They detail how the flaw exposed proprietary AI data. Issue highlights the importance of analyzing complex application logic. https://api.cyfluencer.com/s/hack-the-ai-brain-uncovering-an-account-takeover-vulnerability-in-langsmith-25853
Post summary
A critical account takeover CVE (CVE-2026-25750) affecting the LangSmith platform is reported, noting proprietary AI data exposure, but lacking PoC, exploit code, patch information, or active exploitation details.
0000199
9.1K followersView on X
InfoSecSherpa 🏔️@InfoSecSherpa211d
Disclosure
"Hack the #AI Brain: Uncovering an Account Takeover Vulnerability in LangSmith" Miggo Security discovered a critical account takeover vulnerability (CVE-2026-25750) in LangSmith. Learn how this flaw exposed proprietary AI data. https://api.cyfluencer.com/s/hack-the-ai-brain-uncovering-an-account-takeover-vulnerability-in-langsmith-25858 https://t.co/Rd5i56Lcnh
Post summary
The tweet announces the discovery of CVE-2026-25750 as an account takeover flaw in LangSmith, but provides no PoC, exploit, or mitigation details.
00010280
51.6K followersView on X
Infoflowcloud@infoflowcloud219d
Disclosure
🚨*CVE* CVE-2026-25750 Langchain Helm Charts are Helm charts for deploying Langchain applications on Kubernetes. Prior to langchain-ai/helm version 0.12.71, a URL parameter injection vulner… https://www.cve.org/CVERecord?id=CVE-2026-25750 ----- Traducción: CVE-2026-25750 Lan… http://infoflow.cloud`
Post summary
The post announces a CVE-2026-25750 URL parameter injection flaw in older Langchain Helm charts, citing the version affected and linking to the official CVE record.
0001036
55 followersView on X
CVE@CVEnew219d
Patch
CVE-2026-25750 Langchain Helm Charts are Helm charts for deploying Langchain applications on Kubernetes. Prior to langchain-ai/helm version 0.12.71, a URL parameter injection vulner… https://www.cve.org/CVERecord?id=CVE-2026-25750
Post summary
The announcement highlights a URL parameter injection flaw in older Langchain Helm charts, which was addressed in the 0.12.71 release, providing a clear patch reference.
00010180
56.6K followersView on X
The Daily Tech Feed@dailytechonx208d
Patch
Critical vulnerability CVE-2026-25750 in LangSmith exposes users to account takeovers. Immediate updates required to safeguard AI environments. Link: https://thedailytechfeed.com/critical-langsmith-vulnerability-risking-account-takeovers-urgent-update-required/ #Security #Vulnerability #Update #AI #Technology #Software #Privacy #Hacking #Threat #Patch #Bug #Exploit #Network #System #IT #Ransomware #Cyber #Malware #Breach #Defense
Post summary
The tweet warns of CVE‑2026‑25750, a critical LangSmith flaw that could lead to account takeovers, and urges users to apply updates immediately to protect their AI environments.
0000012
264 followersView on X
ThreatSynop@ThreatSynop210d
Disclosure
🚨 Critical LangSmith Flaw Could Let Attackers Hijack AI Monitoring Accounts Through Session Token Theft Cyber Press reports that CVE-2026-25750 in LangSmith let attacker-controlled base URLs capture authenticated session credentials, creating a short window for silent account takeover and access to sensitive AI trace data. This matters because compromised LangSmith accounts could expose proprietary prompts, internal query results, customer data, and other high-value enterprise AI telemetry. 🎯 Target: Global/Enterprise AI Environments #️⃣ Category: #Vulnerability #AI_Threats #BlueTeam #CyberIntel 🔗 URL: https://cyberpress.org/critical-langsmith-vulnerability/
Post summary
Cyber Press reports CVE-2026-25750 as a session-token theft flaw in LangSmith that could allow attackers to hijack accounts, but no PoC, exploit, patch, or evidence of active exploitation is provided.
0000056
286 followersView on X
ThreatSynop@ThreatSynop210d
Disclosure
🚨 Critical LangSmith Flaw Could Let Attackers Hijack AI Monitoring Accounts Through Session Token Theft Cyber Security News reports that CVE-2026-25750 in LangSmith allowed attacker-controlled base URLs to receive authenticated API requests and session credentials, creating a five-minute window for silent account takeover and access to sensitive AI trace data. This matters because compromised LangSmith accounts could expose proprietary prompts, internal database outputs, customer data, and other high-value enterprise AI telemetry. 🎯 Target: Global/Enterprise AI Environments #️⃣ Category: #Vulnerability #AI_Threats #BlueTeam #CyberIntel 🔗 URL: https://cybersecuritynews.com/critical-langsmith-account-takeover-vulnerability/
Post summary
The post announces CVE‑2026‑25750 in LangSmith, explaining how attacker-controlled URLs can hijack session tokens for a limited takeover window, but provides no PoC, patch, or evidence of active exploitation.
0000053
286 followersView on X

CPE platform detail1 entries

Part:

Vendor:

Product:

Version:

Target SW:

Target HW:

1 of 1 entries

Part	Vendor	Product	Version	Target SW	Target HW
App	langchain	langsmith	-	kubernetes	-

Explore more

Back to dashboard langchain CVEs langsmith vulnerabilities

What researchers are saying

CVE-2026-25750

Sekurak[verified]@SekurakPatch

CVE‑2026‑25750 in LangSmith permits session hijacking via an unsanitized baseURL parameter; the issue was fixed in version 0.12.71, and users—especially self‑hosted deployments—are urged to update immediately.

Audrey Renée Bentley[verified]@BentleyAudreyDisclosure

The excerpt announces the discovery of a critical account takeover flaw (CVE‑2026‑25750) in LangSmith, highlighting its potential to expose proprietary AI data.

The Daily Tech Feed[verified]@dailytechonxPatch

The tweet warns of CVE‑2026‑25750, a critical LangSmith flaw that could lead to account takeovers, and urges users to apply updates immediately to protect their AI environments.

ThreatSynop[verified]@ThreatSynopDisclosure

Cyber Press reports CVE-2026-25750 as a session-token theft flaw in LangSmith that could allow attackers to hijack accounts, but no PoC, exploit, patch, or evidence of active exploitation is provided.

ThreatSynop[verified]@ThreatSynopDisclosure

The post announces CVE‑2026‑25750 in LangSmith, explaining how attacker-controlled URLs can hijack session tokens for a limited takeover window, but provides no PoC, patch, or evidence of active exploitation.

iototsecnews@iototsecnewsPatch

The post explains CVE-2026-25750 in LangSmith, detailing the flaw and the vendor’s fix that enforces a whitelist of trusted origins, but it offers no PoC, exploit code, or evidence of active exploitation.

SecAlerts@SecAlertsCoPatch

The tweet announces a high‑severity vulnerability in LangSmith (CVE‑2026‑25750) that could allow token theft and account takeover, and directs readers to SecAlerts for fix information.

Ben Rothke@benrothkeDisclosure

A critical account takeover CVE (CVE-2026-25750) affecting the LangSmith platform is reported, noting proprietary AI data exposure, but lacking PoC, exploit code, patch information, or active exploitation details.