CVE-2026-77521Patch

MEDIUMCVSS 10.0 · CRITICAL

Exploit discussion active in current signal (1 latest mentions)

Immediate actions

Patch affected systems immediately
Hunt for exploitation attempts and persistence artifacts
Increase monitoring for publicly documented tradecraft

Recommended action window: High priority (within 72h)

NVD description

MaxKB is an open-source AI assistant for enterprise. Prior to version 2.10.5-lts, assistants with a tool, MCP tool, skill, or sub-application use SandboxShellBackend, which exposes an execute shell tool without excluding it and omits execute from interrupt_on, so human approval is not required. Untrusted chat or ingested content can therefore cause command execution; source deployments with MAXKB_SANDBOX disabled run commands directly as the application user, while the official root container's string-based gosu wrapper allowed shell metacharacters to execute outside the intended sandbox. This issue is fixed in version 2.10.5-lts.

Back to dashboard

4.0/ 10 priority

Sources & remediation

Other references

Weakness type (CWE)

CWE-78CWE-250CWE-749

Priority

MEDIUM

Exploitation

NONE

PoC

YES

Patch

AVAILABLE

Momentum

STABLE

Threat summary

Public PoC and exploit tooling are both present
Patch or workaround signal is available
11 mentions across 6 observed days
Momentum state: stable

What's happening

Exploit tool or code specified in 1 signal
PoC mentioned or linked in 4 signals
Patch or workaround mentioned in 4 signals
Technical details provided in 9 signals
Disclosure: 3 classified signals
Peaked 5d ago at 4 mentions (2026-09-22); latest day: 1
11 total mentions across 6 days

Deep dive

Activity timeline11 mentions / 6d

Signal classification5 categories

Patch

436.4%

Disclosure

327.3%

PoC

218.2%

General

19.1%

Exploit

19.1%

Referenced assets6 URLs

By indicator

Classification over time

Date	Total	Labels

2026-09-22	4	Disclosure1General1Patch1PoC1
2026-09-24	1	Disclosure1
2026-09-26	1	Patch1
2026-10-15	1	Patch1
2026-10-17	3	Disclosure1Exploit1PoC1
2026-10-26	1	Patch1

Full discourse11 posts

Rıdvan Yağlı@ridvanyagli12d
Patch
🔴 Hosting kontrol panellerle ilgili son haftalarda bilinen güvenlik açıkları: ◼️ CyberPanel - CVE-2026-71965 ve CVE-2026-71966 (CVSS ~8.8 High) ● Authenticated RCE / command injection (remote backup özelliği). Root SSH erişimi veya komut çalıştırma. 26 Ekim'de DFIR Lab tarafından paylaşıldı. Commit eca0c3c ile fixlenmiş. ◼️ aaPanel - CVE-2026-113506 (CVSS 9.9 / 8.6) ● File Manager'da authenticated OS command injection (ExecShell). Public PoC var. Vendor authentication gerektiriyor diyor. 20 Ekim civarı yayınlandı. aaPanel ayrıca PHP2Shell için geçici hardening duyurusu yaptı. ◼️ Froxlor - CVE-2026-113678 (CVSS 9.9) ● Authenticated hosting tenant'ların symlink ile root cron chown'unu ele geçirmesi -> root kontrol. 2.3.15+ sürümüne yükseltin uyarısı var. ◼️ 1Panel - CVE-2026-77521 (CVSS 10) ● ExploitGrid gibi hesaplar "lab environment live" olarak listeledi. Son günlerde trending exploits listelerinde görünüyor.
Post summary
The text discloses multiple CVEs in hosting control panels with technical details and CVSS scores, while also referencing fixes such as commits, version upgrades, and hardening advisories; a public PoC is noted for aaPanel but no active exploitation or named exploit tool is confirmed.
10024102.0K
2.9K followersView on X
Welsh ICP Conviction 🏴󠁧󠁢󠁷󠁬󠁳󠁿🏉@ICPLEGEND196646d
Disclosure
🚨 internet-computer:native — AI AGENTS ARE GETTING POWERFUL ENOUGH TO TURN UNTRUSTED DATA INTO SYSTEM COMMANDS. 🤖🔐♾️ MaxKB has disclosed a CVSS 10.0 vulnerability, CVE-2026-77521. The reported chain is alarming: 📄 Untrusted content / chat → 🤖 AI agent interprets instructions → 🛠️ Tool access → 💻 Shell command execution → ⚠️ Missing human-approval boundary This is the fundamental AI security problem: CONTENT IS NO LONGER JUST DATA. With autonomous agents, content can become INSTRUCTIONS — and instructions can become ACTIONS. This is exactly why the architecture behind internet-computer:native and @dfinity is becoming increasingly relevant. ♾️ Tamper-resistant execution 🔐 Cryptographic authorization 🛡️ Protocol-level security & resilience ☁️ Sovereign Cloud Engines 🤖 AIware with enforceable guardrails ⚙️ Reduced dependence on conventional server stacks internet-computer:native cannot eliminate prompt injection or bad application logic. But the AI era needs infrastructure designed around the assumption that autonomous software WILL hold real authority. AI is accelerating. THE SECURITY ARCHITECTURE UNDERNEATH IT MUST ACCELERATE TOO. 🛡️♾️🔥 @dfinity @OccupyDemocrats #ICP #InternetComputer #AISecurity #CyberSecurity #AIware #CloudEngines ♾️ Support my internet-computer:native content: 1e672d038cebc619d93186418fa98f6499dbdb9cfdfac54f366c61a4a4ee4362
Post summary
MaxKB disclosed a CVSS 10.0 vulnerability (CVE-2026-77521) where untrusted chat content can be turned into shell commands by autonomous AI agents due to missing human‑approval controls, emphasizing the need for security‑by‑design AI infrastructure.
00080235
1.8K followersView on X
ExploitGrid@exploitgrid21d
Disclosure
ExploitGrid Daily Digest 🚨 Top Exploits: CVE-2025-68613 (CVSS: 10) n8n-io CVE-2026-110958 (CVSS: 10) wavlink CVE-2026-77521 (CVSS: 10) 1Panel-dev CVE-2026-20329 (CVSS: 9.9) Cisco CVE-2026-104711 (CVSS: 9.8) Apache Softw... ..🧵👇
Post summary
A concise daily digest listing top CVEs with their CVSS scores and affected products; the text serves to announce and disclose vulnerability information without providing PoCs, exploit tools, patches, or active exploitation evidence.
11020223
408 followersView on X
VulnTracker@vuln_tracker46d
Patch
One AI assistant, two separate paths to root (CVSS 10.0 and 9.1). CVE-2026-77521: MaxKB's tool and MCP agents expose a shell-execute function without human approval, so untrusted chat content alone can trigger command execution. CVE-2026-79916: A workspace member can inject control characters into AWS Bedrock credential fields, planting a malicious profile that runs an attacker's command as root. Both fixed in MaxKB 2.10.5-lts. Update now. Details: http://vulntracker.io/cves/CVE-2026-77521 #MaxKB #CVE #AI #PromptInjection #InfoSec #CyberSecurity
Post summary
The tweet discloses two MaxKB CVEs (CVE‑2026‑77521 and CVE‑2026‑79916) involving command execution and credential injection, provides CVSS scores, and states they are both patched in MaxKB 2.10.5‑lts, urging an update.
10030177
832 followersView on X
ExploitGrid@exploitgrid21d
PoC
💀 CRITICAL Exploits Trending ├ CVE-2026-110958 · Wavlink · PoC live, 1 day after disclosure (10) ├ CVE-2026-77521 · 1Panel · Lab environment live (10) ├ CVE-2025-68613 · n8n · CTF environment live (10) ├ CVE-2026-20329 · Cisco · Predicted-exploit entry (9.9)
Post summary
The text is a trending exploits list primarily highlighting CVEs with live Proof of Concepts or lab/CTF environments available, with the core takeaway being PoC availability rather than active exploitation or named exploit tools.
1000085
408 followersView on X
ExploitGrid@exploitgrid21d
Exploit
[EXPLOIT] CVE-2026-77521 [CRITICAL/PoC] CVSS: 10 | Vendor: #1Paneldev cve-2026-77521-lab 🔗 https://exploitgrid.net/exploits/3dc33498-10ff-4487-8990-0e8a55a6cdc9
Post summary
The text announces CVE-2026-77521 as a critical vulnerability (CVSS 10) and prominently features an [EXPLOIT] tag with a direct link to exploit code on exploitgrid.net, making exploit availability the primary takeaway.
1000039
406 followersView on X
SecAlerts@SecAlertsCo42d
Patch
🤖 MaxKB CVE-2026-77521 is CVSS 10. Prompt-injectable agents with tool/MCP/skill access abuse SandboxShellBackend to execute OS commands. No auth required, full C/I/A impact. Patch to v2.10.5-lts now. #cybersecurity #ciso #vulnerabilities #msp https://secalerts.co/vulnerability/CVE-2026-77521?utm_campaign=x https://t.co/MMlFAtBFhc
Post summary
The tweet discloses CVE-2026-77521 (CVSS 10), explaining how prompt-injectable agents can abuse SandboxShellBackend to execute OS commands without authentication, and urges users to patch to version v2.10.5-lts.
0010097
894 followersView on X
ExploitGrid@exploitgrid46d
PoC
[CVE] CVE-2026-77521 [HIGH PRIORITY] #MaxKB: Prompt-injectable agent can lead to command execution 🔗 https://exploitgrid.net/cve/CVE-2026-77521
Post summary
A high-priority CVE in MaxKB is disclosed involving prompt injection leading to command execution, with a link to exploitgrid.net implying PoC/exploit details are available at the referenced resource.
1000033
203 followersView on X
ExploitGrid@exploitgrid46d
General
🛡️ #ExploitGrid Daily #Threat Digest Top Vulnerabilities (CVEs) of the day CVE-2026-77521 CVE-2026-79920 CVE-2026-82187 CVE-2026-85751 CVE-2026-94301 ..🧵👇
Post summary
The tweet simply enumerates a list of CVEs as daily top vulnerabilities without additional context, remediation, or exploitation details.
1000055
203 followersView on X
Hugo | DevOps | Cybersecurity 🇱🇻@HugoValters23d
Patch
CVE-2026-77521 MaxKB CVSS 10: shell execute tool exposed without approval, untrusted chat input leads to RCE. Patch is under review, so isolate or disable MAXKB_SANDBOX now. https://www.valtersit.com/cve/CVE-2026-77521 #CVE #infosec #MaxKB #MaxKB #CVEALERT #CVE #infosec #Linux #XSS #valtersit #snippet #SysAdmin #cybersecurity #devsecops #devops #developer #100daysofcode #git #github #gitlab #redteam #blueteam #ethicalhacker #cybersecurityawareness #cybersecuritynews #cybersecuritytips #python #hacker #kali #ubuntu #debian #docker
Post summary
The text discloses CVE-2026-77521, a CVSS 10 RCE vulnerability in MaxKB caused by an exposed shell execute tool, and provides a workaround (disable MAXKB_SANDBOX) while noting a patch is under review.
0000043
1.1K followersView on X
Cybersecurity News DE@cybsecuritynews44d
Disclosure
#schwachstellen MaxKB: Kritische Lücken erlauben Remote-Code-Ausführung bis Root vor Version 2.10.5-lts #awsbedrock #cve202677521 #cve202679916 #maxkb https://cybersecurity-news.de/maxkb-kritische-schwachstellen-cve-2026-77521-cve-2026-79916
Post summary
The tweet discloses critical remote code execution vulnerabilities in MaxKB affecting versions before 2.10.5-lts, referencing two CVE IDs, without mentioning exploits, patches, or active exploitation.
0000029
15 followersView on X

Explore more

Back to dashboard

What researchers are saying

CVE-2026-77521

Rıdvan Yağlı[verified]@ridvanyagliPatch

The text discloses multiple CVEs in hosting control panels with technical details and CVSS scores, while also referencing fixes such as commits, version upgrades, and hardening advisories; a public PoC is noted for aaPanel but no active exploitation or named exploit tool is confirmed.

Welsh ICP Conviction 🏴󠁧󠁢󠁷󠁬󠁳󠁿🏉[verified]@ICPLEGEND1966Disclosure

MaxKB disclosed a CVSS 10.0 vulnerability (CVE-2026-77521) where untrusted chat content can be turned into shell commands by autonomous AI agents due to missing human‑approval controls, emphasizing the need for security‑by‑design AI infrastructure.

VulnTracker[verified]@vuln_trackerPatch

The tweet discloses two MaxKB CVEs (CVE‑2026‑77521 and CVE‑2026‑79916) involving command execution and credential injection, provides CVSS scores, and states they are both patched in MaxKB 2.10.5‑lts, urging an update.

Hugo | DevOps | Cybersecurity 🇱🇻[verified]@HugoValtersPatch

The text discloses CVE-2026-77521, a CVSS 10 RCE vulnerability in MaxKB caused by an exposed shell execute tool, and provides a workaround (disable MAXKB_SANDBOX) while noting a patch is under review.

ExploitGrid@exploitgridDisclosure

A concise daily digest listing top CVEs with their CVSS scores and affected products; the text serves to announce and disclose vulnerability information without providing PoCs, exploit tools, patches, or active exploitation evidence.

ExploitGrid@exploitgridPoC

The text is a trending exploits list primarily highlighting CVEs with live Proof of Concepts or lab/CTF environments available, with the core takeaway being PoC availability rather than active exploitation or named exploit tools.

ExploitGrid@exploitgridExploit

The text announces CVE-2026-77521 as a critical vulnerability (CVSS 10) and prominently features an [EXPLOIT] tag with a direct link to exploit code on exploitgrid.net, making exploit availability the primary takeaway.

SecAlerts@SecAlertsCoPatch

The tweet discloses CVE-2026-77521 (CVSS 10), explaining how prompt-injectable agents can abuse SandboxShellBackend to execute OS commands without authentication, and urges users to patch to version v2.10.5-lts.