CVE-2026-21852

Published Jan 21, 2026

Last updated 2 months ago

Overview

Description
Claude Code is an agentic coding tool. Prior to version 2.0.65, vulnerability in Claude Code's project-load flow allowed malicious repositories to exfiltrate data including Anthropic API keys before users confirmed trust. An attacker-controlled repository could include a settings file that sets ANTHROPIC_BASE_URL to an attacker-controlled endpoint and when the repository was opened, Claude Code would read the configuration and immediately issue API requests before showing the trust prompt, potentially leaking the user's API keys. Users on standard Claude Code auto-update have received this fix already. Users performing manual updates are advised to update to version 2.0.65, which contains a patch, or to the latest version.
Source
security-advisories@github.com
NVD status
Analyzed
Products
claude_code

Risk scores

CVSS 4.0

Type
Secondary
Base score
5.3
Impact score
-
Exploitability score
-
Vector string
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X
Severity
MEDIUM

CVSS 3.1

Type
Primary
Base score
7.5
Impact score
3.6
Exploitability score
3.9
Vector string
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
Severity
HIGH

Weaknesses

security-advisories@github.com
CWE-522

Social media

Hype score is a measure of social media activity compared against trending CVEs from the past 12 months. Max score 100.

Hype score

4

  1. Caught in the Hook: RCE and API Token Exfiltration Through Claude Code Project Files | CVE-2025-59536 | CVE-2026-21852 - Aviv Donenfeld and Oded Vanunu https://t.co/ne7IicPoHh

    @pentest_swissky

    11 Apr 2026

    1686 Impressions

    7 Retweets

    17 Likes

    13 Bookmarks

    0 Replies

    0 Quotes

  2. 在陌生 repo 跑 Claude Code,API 金鑰可能已洩漏。 Check Point 披露 CVE-2025-59536(任意命令執行)+ CVE-2026-21852(密鑰外滲),攻擊鏈全程無感知,Anthropic 未有修復時間表。 vibe coding 用戶每天都咁做——邊緣場景定日

    @TechPulseHK

    2 Apr 2026

    112 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  3. 1️⃣ About Claude Code vulnerabilities (CVE-2025-59536, CVE-2026-21852) : the attack surface no longer exists at the code execution layer alone. It has migrated upward into the configuration and initialization layers that govern how AI assistants interact with infrastructure b

    @francescofaenzi

    30 Mar 2026

    141 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    1 Reply

    1 Quote

  4. 4 Agent Security Breaches That Should Change How You Think About API Keys Shai-Hulud, SANDWORM_MODE, OpenClaw, and CVE-2026-21852 — the 2025-2026 incidents that proved AI agents need their own authorization layer, not shared secrets. https://t.co/JI0XipSONE

    @mishrak_sanjeev

    28 Mar 2026

    129 Impressions

    0 Retweets

    5 Likes

    2 Bookmarks

    0 Replies

    0 Quotes

  5. 毎日使ってるClaude Codeに重大な脆弱性が2件見つかった(修正済み)。 CVE-2025-59536: プロジェクトファイル経由のRCE CVE-2026-21852: ANTHROPIC_BASE_URLの上書きによるAPIキー漏洩 攻撃シナリオがリアル。 悪意あるリポ

    @shun_aidev

    19 Mar 2026

    249 Impressions

    0 Retweets

    2 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  6. Recent #ClaudeCode vulnerabilities (CVE-2025-59536, CVE-2026-21852) show why the AI supply chain begins with the automation layer. Configuration files like .claude/settings.json are now part of the execution layer. Authority must be deterministic. 🛡️🦾 https://t.co/PGucc61

    @PermissionPrtcl

    17 Mar 2026

    150 Impressions

    0 Retweets

    5 Likes

    0 Bookmarks

    1 Reply

    0 Quotes

  7. #Cybernews 🚨💻 Las #Vulnerabilidades ⚠️, registradas como CVE-2025-59536 🆔 y CVE-2026-21852 🆔, podían activarse simplemente al #Clonar 🔁 y abrir 📂 un proyecto no confiable 🚫, sin necesidad de ejecutar código explícito 🧩 ni realizar acciones adicional

    @totalcybersec

    10 Mar 2026

    219 Impressions

    2 Retweets

    6 Likes

    0 Bookmarks

    1 Reply

    0 Quotes

  8. 🚨 CYBERDUDEBIVASH SENTINEL APEX ALERT 🚨 Threat: Caught in the Hook: RCE and API Token Exfiltration Through Claude Code Project Files | CVE-2025-59536 | CVE-2026-21852 Intel Report: https://t.co/1U7gyRBUzM

    @cyberbivash

    9 Mar 2026

    97 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  9. check point found 2 flaws in claude code — anthropic's AI dev tool. CVE-2025-59536: open a project → code runs before you click "trust." CVE-2026-21852: repo configs silently redirect your API keys to the attacker. clone the wrong repo. your AI tool is the backdoor.

    @The_Agent_Econ

    8 Mar 2026

    104 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  10. 🚨 CYBERDUDEBIVASH SENTINEL APEX ALERT 🚨 Threat: Caught in the Hook: RCE and API Token Exfiltration Through Claude Code Project Files | CVE-2025-59536 | CVE-2026-21852 Intel Report: https://t.co/tL4HD8m3Hz

    @cyberbivash

    3 Mar 2026

    109 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  11. Claude Codeの設定ファイルが攻撃面に不正リポジトリでRCEとAPIキー窃取が成立した脆弱性(CVE-2025-59536/CVE-2026-21852) https://t.co/Vs50pJ5SIK #セキュリティ対策Lab #セキュリティ #Security #CybersecurityNews #AINews

    @securityLab_jp

    3 Mar 2026

    183 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  12. Vulnerabilities (CVE-2025-59536, CVE-2026-21852) in Anthropic Claude Code https://t.co/k7E25uyOfG

    @ninp0

    2 Mar 2026

    48 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  13. Claude Code มีช่องโหว่ เปิดทางแฮกเกอร์รันคำสั่ง-ขโมย API Key ได้เงียบๆ https://t.co/8UxH8udJWf CVE-2025-59536, CVE-2026-21852

    @ohmohm

    2 Mar 2026

    50 Impressions

    0 Retweets

    1 Like

    0 Bookmarks

    0 Replies

    0 Quotes

  14. AI dev tool security alert. @claudeai Code vulnerabilities enabled: • Remote Code Execution • MCP consent bypass (CVE-2025-59536) • API key exfiltration (CVE-2026-21852) Reported by Check Point Research. Fully patched by Anthropic. Config files = potential execution vectors

    @TechNadu

    28 Feb 2026

    188 Impressions

    0 Retweets

    1 Like

    0 Bookmarks

    0 Replies

    0 Quotes

  15. Report #2026-02-28-01: Claude Code flaws (CVE-2025-59536, CVE-2026-21852) enabled RCE + API token exfiltration via untrusted project files. Impact: HIGH. Source: https://t.co/rX49vfvl2V

    @elagentecapital

    28 Feb 2026

    14 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    1 Reply

    0 Quotes

  16. 「リポジトリをクローンして開いただけで APIキーが盗まれる」 セキュリティ企業Check Pointが Claude Codeに重大な脆弱性2件を発見・報告。 CVE-2025-59536 CVE-2026-21852 2件とも公開前にAnthropicが修正済みです🔐 ど

    @Claudia_AiLab

    27 Feb 2026

    109 Impressions

    0 Retweets

    1 Like

    0 Bookmarks

    1 Reply

    0 Quotes

  17. ⚠️Check PointがClaude Codeの重大脆弱性を公開 The Hacker Newsでも報道されました 内容はかなり深刻で ・悪意あるリポジトリを開くだけ ・RCE(Remote Code Execution=外部から任意コード実行)可能 ・APIキー盗

    @onumaro92

    26 Feb 2026

    210 Impressions

    0 Retweets

    2 Likes

    0 Bookmarks

    1 Reply

    0 Quotes

  18. Check Point disclosed critical Claude Code vulnerabilities yesterday (CVE-2025-59536, CVE-2026-21852). Three attack vectors, all execution before trust dialogs. RCE via hooks: Malicious .claude/settings.json executes shell commands on SessionStart. Clone poisoned repo, run

    @ManfredMancxx

    26 Feb 2026

    52 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    0 Replies

    0 Quotes

  19. 🚨 Critical RCE in Code: How Attackers Can Hijack #AI Assistants and Steal API Keys (#CVE-2025-59536 & #CVE-2026-21852) + Video https://t.co/aOkUZKfrk5 Educational Purposes!

    @UndercodeUpdate

    26 Feb 2026

    49 Impressions

    0 Retweets

    1 Like

    0 Bookmarks

    0 Replies

    0 Quotes

  20. I hacked Claude Code! It turns out "agentic" is just a fancy new way to get a shell. I achieved full RCE and hijacked organization API keys. CVE-2025-59536 | CVE-2026-21852 https://t.co/GymKzaM1wp #ai #Claude

    @Od3dV

    26 Feb 2026

    60136 Impressions

    102 Retweets

    470 Likes

    334 Bookmarks

    6 Replies

    13 Quotes

  21. Critical vulnerabilities, CVE-2025-59536 and CVE-2026-21852, in Anthropic’s Claude Code enabled remote code execution and API key theft through malicious repository-level configuration files, triggered simply by cloning and opening an untrusted project https://t.co/nXlrDqdhgK h

    @blackorbird

    26 Feb 2026

    2078 Impressions

    9 Retweets

    21 Likes

    8 Bookmarks

    0 Replies

    0 Quotes

  22. What dropped today (while these clowns are still selling unsecured garbage bots that get your account nuked): • Claude Code Config Bypass/CVE-2025-59536 + CVE-2026-21852 lets attackers RCE your dev box and steal API keys just by cloning a poisoned repo — disclosed Feb 25, 20

    @Double00Kevin

    26 Feb 2026

    62 Impressions

    0 Retweets

    1 Like

    0 Bookmarks

    0 Replies

    0 Quotes

  23. Anthropic社のAI「Claude Code」において、リモートでコードが実行される可能性のある深刻な脆弱性が発見されました。この問題は、CVE-2025-59536およびCVE-2026-21852として追跡されています。

    @omomuki_tech

    26 Feb 2026

    75 Impressions

    0 Retweets

    1 Like

    0 Bookmarks

    0 Replies

    0 Quotes

  24. Claude Code Flaws Allow Remote Code Execution and API Key Exfiltration - https://t.co/yCtHWMfO00 • Critical vulnerabilities, CVE-2025-59536 and CVE-2026-21852, in Anthropic’s Claude Code enabled remote code execution and API key theft through malicious repository-level https

    @AISecHub

    26 Feb 2026

    1660 Impressions

    3 Retweets

    26 Likes

    13 Bookmarks

    2 Replies

    0 Quotes

  25. Check Point found CVE-2025-59536 and CVE-2026-21852 in Claude Code allow remote code execution and API key theft via untrusted repository configurations, reachable by simply cloning and opening a project. They warn that built-in hooks and env vars could … https://t.co/nEKaQHLG0

    @Cyber_O51NT

    26 Feb 2026

    943 Impressions

    12 Retweets

    15 Likes

    1 Bookmark

    2 Replies

    0 Quotes

  26. Researchers disclose critical flaws in Anthropic's Claude Code enabling remote code execution and API key theft via untrusted repositories, tracked as CVE-2025-59536 and CVE-2026-21852. #AIsecurity https://t.co/49GjRjE368

    @threatcluster

    26 Feb 2026

    64 Impressions

    0 Retweets

    0 Likes

    0 Bookmarks

    1 Reply

    0 Quotes

  27. Claude Codeに遠隔コード実行及びAPIキー窃取の脆弱性。Check Point社報告。CVE-2025-59536及びCVE-2026-21852。信頼されないリポジトリをクローンして開くことで、悪意ある設定ファイルから発動。処理に際し明示的な認

    @__kokumoto

    25 Feb 2026

    1147 Impressions

    3 Retweets

    3 Likes

    5 Bookmarks

    0 Replies

    0 Quotes

  28. Caught in the Hook: RCE and API Token Exfiltration Through Claude Code Project Files | CVE-2025-59536 | CVE-2026-21852 https://t.co/fTjad1Bg5x

    @Dinosn

    25 Feb 2026

    1281 Impressions

    2 Retweets

    1 Like

    4 Bookmarks

    0 Replies

    0 Quotes

  29. Check Point | Caught in the Hook: RCE and API Token Exfiltration Through Claude Code Project Files | CVE-2025-59536 | CVE-2026-21852 https://t.co/4ymK4vGUTN

    @StopMalvertisin

    25 Feb 2026

    349 Impressions

    1 Retweet

    5 Likes

    1 Bookmark

    0 Replies

    0 Quotes

Configurations