CVE-2026-44338Active Exploitation(praison / praisonai)

HIGHCVSS 7.3 · HIGH

Exploitation observed; activity peaked at 21 mentions and remains active

Immediate actions

Patch praison praisonai systems immediately
Assume compromise if assets are exposed
Hunt for exploitation attempts and persistence artifacts
Increase monitoring for publicly documented tradecraft

Recommended action window: Immediate (within 24h)

NVD description

PraisonAI is a multi-agent teams system. From version 2.5.6 to before version 4.6.34, PraisonAI ships a legacy Flask API server with authentication disabled by default. When that server is used, any caller that can reach it can access /agents and trigger the configured agents.yaml workflow through /chat without providing a token. This issue has been patched in version 4.6.34.

praison CVEs Back to dashboard

6.5/ 10 priority

Sources & remediation

Exploit / PoC references

Weakness type (CWE)

CWE-306CWE-668CWE-1188

Priority

HIGH

Exploitation

ACTIVE

PoC

YES

Patch

AVAILABLE

Momentum

DECLINING

Are you affected?

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

praisonai

Threat summary

Active exploitation appears in 49 classified signals
Public PoC is present in monitored signal
Patch or workaround signal is available
62 mentions across 19 observed days

What's happening

Active exploitation reported across 49 signals
PoC mentioned or linked in 2 signals
Patch or workaround mentioned in 18 signals
Technical details provided in 43 signals
General: 5 classified signals
Peaked 15d ago at 21 mentions (2026-05-14); latest day: 1
62 total mentions across 19 days

Affected systems

Vendors

praison

Products

praisonai

Deep dive

Activity timeline62 mentions / 19d

Signal classification4 categories

Active Exploitation

4775.8%

Patch

69.7%

General

58.1%

Disclosure

46.5%

Referenced assets30 URLs

By indicator

Classification over time

Date	Total	Labels

2026-05-08	1	Disclosure1
2026-05-12	1	Active Exploitation1
2026-05-13	6	Active Exploitation6
2026-05-14	21	Active Exploitation17Disclosure1Patch3
2026-05-15	6	Active Exploitation6
2026-05-16	5	Active Exploitation4General1
2026-05-17	1	Active Exploitation1
2026-05-18	5	Active Exploitation4Disclosure1
2026-05-19	1	Patch1
2026-05-21	2	Active Exploitation2
2026-05-22	2	Active Exploitation1Patch1
2026-05-23	1	Active Exploitation1
2026-05-26	1	Patch1
2026-05-27	3	Active Exploitation1General2
2026-06-11	1	Active Exploitation1
2026-06-12	1	Active Exploitation1
2026-07-21	2	General2
2026-07-25	1	Active Exploitation1
2026-08-20	1	Disclosure1

Full discourse20 posts

The Hacker News@TheHackersNews149d
Active Exploitation
🚨 Threat actors targeted PraisonAI CVE-2026-44338, an authentication bypass vulnerability, within hours of disclosure. The flaw affects versions 2.5.6–4.6.33 and can expose the /agents endpoint without authorization. Read the full report: https://thehackernews.com/2026/05/praisonai-cve-2026-44338-auth-bypass.html
Post summary
The report confirms that threat actors actively exploited CVE-2026-44338 within hours of its disclosure, highlighting an authentication bypass that exposes the /agents endpoint in certain PraisonAI versions.
3813849.0K
1.9M followersView on X
Gray Hats@the_yellow_fall150d
Active Exploitation
Sysdig warns: PraisonAI (CVE-2026-44338) exploited in under 4 hours. Attackers bypassed auth to hijack agents and drain API quotas. Update to 4.6.34 now! #PraisonAI #CyberSecurity #InfoSec #AISecurity #VulnerabilityAlert #CVE202644338 #ExploitAlert https://securityonline.info/praisonai-vulnerability-cve-2026-44338-exploited-in-the-wild/ https://t.co/UzGuHpfU6R
Post summary
Sysdig warns that PraisonAI CVE‑2026‑44338 was actively exploited within 4 hours, with attackers bypassing authentication to hijack agents and drain API quotas; upgrading to 4.6.34 is recommended.
04072566
12.5K followersView on X
Sysdig@sysdig149d
Active Exploitation
3 hours. 44 minutes. That's how long it took for active scanning to start after CVE-2026-44338 dropped. Advisory-to-exploitation windows are now measured in single-digit hours. Every AI project is a target. Full research from Sysdig TRT: https://okt.to/wiS9LD #CloudSecurity #ThreatResearch #AIAgents
Post summary
The message highlights that CVE-2026-44338 began being actively exploited only about 3.5 hours after disclosure, underscoring the rapid exploitation risk for AI-focused projects.
12081485
10.3K followersView on X
TodayInCyber@TodayInCyberIO148d
Active Exploitation
🔴 CRITICAL | Threat actors have been observed exploiting CVE-2026-44338, a critical authentication bypass vulnerability in the PraisonAI open-source multi-agent orchestration framework, within four hours of i… #vulnerability #CVE202644338 #cybersecurity https://todayincyber.io/feed/vulnerability/article/019e2670-d471-7860-b133-03b65850b166
Post summary
Threat actors have exploited CVE-2026-44338 within hours of its public disclosure; no PoC, exploit code, patch, or false‑positive claim is provided, and the vulnerability is described as a critical authentication bypass.
03050145
10 followersView on X
Gray Hats@the_yellow_fall145d
Active Exploitation
Critical flaw CVE-2026-44338 in PraisonAI exposes autonomous workflows to unauthenticated hijackers within hours of disclosure. Update to v4.6.34 immediately! #PraisonAI #CVE202644338 #AISecurity2026 #InfoSec #BugBounty #DevSecOps #AIAgents https://meterpreter.org/critical-pre-auth-flaw-cve-2026-44338-exploited-to-hijack-autonomous-ai-agents/ https://t.co/1NnNMUVsVs
Post summary
The tweet announces that CVE‑2026‑44338 is being actively exploited within hours of disclosure, urges an immediate patch to version 4.6.34, and includes a link likely pointing to a PoC or exploit details.
01051655
12.5K followersView on X
connect24h@connect24h147d
Active Exploitation
「公開4時間以内に悪用」AIエージェントフレームワーク PraisonAI で認証バイパス脆弱性（CVE-2026-44338、CVSS 7.3）が発覚。開示からわずか4時間未満で攻撃者が動き出した。AIインフラが直接標的になる時代、従来のパッチウィンドウ前提は崩壊 https://is.gd/KTeB08 #AIセキュリティ #CSIRT
Post summary
The report indicates that PraisonAI’s authentication bypass vulnerability (CVE‑2026‑44338, CVSS 7.3) was actively exploited within four hours of disclosure, underscoring the urgency of addressing AI infrastructure security.
01011329
4.6K followersView on X
CyberSecurity88@CSec88149d
Active Exploitation
New AI Security Warning: PraisonAI Vulnerability Targeted in Under 4 Hours Attackers started targeting a newly disclosed AI vulnerability in under 4 hours. #CyberSecurity #AIsecurity #ThreatIntelligence Full Story 👉 https://cybersecurity88.com/news/praisonai-cve-2026-44338-exploited-within-hours-after-public-disclosure/ https://t.co/zg86VEBTeN
Post summary
A newly disclosed PraisonAI vulnerability was reportedly exploited within four hours of its public disclosure, indicating rapid attacker interest and active exploitation.
00030118
546 followersView on X
DFIR Radar@DFIR_Radar150d
Active Exploitation
CVE-2026-44338 exploited in under 4 hours: PraisonAI authentication bypass demonstrates AI-accelerated exploitation becoming the norm. Scanner labeled CVE-Detector/1.0 probing vulnerable endpoints just 3h44m after GitHub advisory. Key technical details: • CVE-2026-44338 (CVSS not specified): Hard-coded AUTH_ENABLED=False in api_server.py exposes GET /agents and POST /chat endpoints • All PyPI releases ≤4.6.33 vulnerable - authentication fails open by design in check_auth() function • Scanner from 146[.]190[.]133[.]49 targeted exact vulnerable paths, confirming successful bypass with 200 OK responses • POST /chat triggers PraisonAI().run() regardless of message content - impact depends on configured agents.yaml workflow Attack methodology: • Two-pass scanning: generic disclosure paths first, then AI-agent specific endpoints • User-Agent "CVE-Detector/1.0" serves as reliable IOC across any targeted infrastructure • Validation-focused approach: enumerate agents, confirm bypass, log as exploitable, move to next target DFIR artifacts: • Network logs showing unauthenticated requests to /agents and /chat without Authorization headers • Python subprocess spawning and unexpected network egress from agent processes post-exploitation • Access attempts to Python fingerprint paths: /pyproject.toml, /poetry.lock, /requirements.txt Deploy WAF rules blocking unauthenticated access to /agents and /chat endpoints. Hunt for User-Agent "CVE-Detector/1. #DFIR_Radar
Post summary
CVE‑2026‑44338 was actively exploited within 4 hours, with an authentication bypass via hard‑coded settings in PraisonAI’s api_server.py, and WAF rules are recommended as a mitigation.
11010252
1.8K followersView on X
Blue Team News@blueteamsec176d
Active Exploitation
PraisonAI CVE-2026-44338 Auth Bypass Targeted Within Hours of Disclosure http://dlvr.it/TTjFwg #PraisonAI #CVE202644338 #AuthBypass #CyberSecurity #Vulnerability https://t.co/CqLL8FGKyG
Post summary
An authorization bypass vulnerability (CVE‑2026‑44338) in PraisonAI was reportedly targeted within hours of its public disclosure, indicating active exploitation in the wild.
000111.8K
56.7K followersView on X
Jim Nitterauer 🇺🇸@JNitterauer135d
Active Exploitation
⏱️ 3hrs 44min. That's how fast CVE-2026-44338 in PraisonAI was exploited after disclosure. AI platforms are the new attack surface. Patch & isolate your AI tools NOW. #AIsecurity #cybersecurity https://thehackernews.com/2026/05/praisonai-cve-2026-44338-auth-bypass.html
Post summary
CVE-2026-44338 in PraisonAI was actively exploited within hours of disclosure, prompting a call for immediate patching and isolation of AI tools.
0002082
8.5K followersView on X
Rock Lambros@rocklambros135d
General
Sysdig research found CVE-2026-44338 in the PraisonAI framework was probed by scanners 3 hours, 44 minutes, and 39 seconds after disclosure. Palo Alto Networks reports 28.3% of CVEs are now exploited within 24 hours. AI-assisted exploit development operates at scale.
Post summary
The tweet notes a vulnerability (CVE‑2026‑44338) was probed shortly after disclosure and cites a general exploitation statistic, without providing technical details, patches, or evidence of active exploitation.
1001074
734 followersView on X
Kwame@kwame_nyx147d
Active Exploitation
Two pieces of AI agent security data published this week: → CVE-2026-44338 dropped on PraisonAI (May 14). The legacy Flask API server hardcodes AUTH_ENABLED = False. Any caller on the network invokes agent workflows. The Hacker News reports the vulnerability was targeted within hours of disclosure. → Akeyless surveyed 400 IT/security leaders running AI agents (May 12). 67% suspect their agents have already accessed unauthorized data. Only 7% believe their controls would prevent a compromised agent from operating. One root cause: the identity layer for AI agents doesn't yet exist as an open standard. Vorim AI is building that layer. cryptographic identity per agent, scoped permissions, signed audit trail.
Post summary
CVE‑2026‑44338 was disclosed and quickly exploited, targeting the Flask API that had AUTH_ENABLED set to false, while no patch or exploit code is publicly referenced.
1001070
60 followersView on X
White Rabbitx 🏴‍☠️@TheRabbitPy148d
Patch
🚨 New auth bypass alert: CVE-2026-44338 in PraisonAI lets attackers potentially skip authentication and reach sensitive functionality. Review your deployment, patch fast, and check exposed AI agent workflows now. #CVE202644338 #PraisonAI #CyberSecurity #AppSec #InfoSec #AIsecurity
Post summary
A new authentication bypass vulnerability (CVE-2026-44338) affects PraisonAI, urging users to review deployments and patch promptly to prevent potential unauthorized access to sensitive AI agent workflows.
1001062
1.3K followersView on X
Nicolas Krassas@Dinosn148d
Active Exploitation
PraisonAI CVE-2026-44338 Auth Bypass Targeted Within Hours of Disclosure https://thehackernews.com/2026/05/praisonai-cve-2026-44338-auth-bypass.html
Post summary
CVE-2026-44338 in PraisAI enables authentication bypass and was reportedly targeted within hours of its public disclosure.
000111.1K
158.6K followersView on X
Vivek | Cybersecurity@VivekIntel148d
Active Exploitation
PraisonAI CVE-2026-44338 Auth Bypass Targeted Within Hours of Disclosure https://thehackernews.com/2026/05/praisonai-cve-2026-44338-auth-bypass.html?m=1
Post summary
The brief text indicates that CVE-2026-44338 is being actively exploited shortly after disclosure, but lacks details on PoC, tools, patches, or technical specifics.
00002265
10.5K followersView on X
SoEmailSecurity@Soemailsecurity148d
Active Exploitation
Threat actors exploited PraisonAI's CVE-2026-44338 auth bypass vulnerability just 4 hours after disclosure, what's your team's plan to respond that quickly to emerging threats? #cybersecurity #vulnerabilitymanagement #PraisonAI
Post summary
Threat actors exploited PraisonAI’s CVE‑2026‑44338, an authentication bypass vulnerability, within hours of its disclosure, indicating active exploitation in the wild.
1001043
60 followersView on X
Upwind Security MDR@UpwindMDR149d
Patch
🚨 High - PraisonAI Legacy API Authentication Bypass (CVE-2026-44338) PraisonAI ships a legacy Flask API server with authentication disabled by default, allowing unauthenticated access to /agents and workflow execution through /chat. A remote attacker can trigger configured agents.yaml workflows without a token, potentially leading to unauthorized agent execution, workflow abuse, and API quota consumption. 👉 Affected: PraisonAI >= 2.5.6 <= 4.6.33 | Fix: Upgrade to 4.6.34 and avoid exposing legacy API deployments without authentication
Post summary
CVE‑2026‑44338 is a high‑severity authentication bypass in PraisonAI’s legacy Flask API, permitting unauthenticated workflow execution; users are advised to upgrade to version 4.6.34 to remediate the vulnerability.
0002097
255 followersView on X
Cybersecurity News Everyday@TweetThreatNews149d
Patch
Sysdig saw scans for PraisonAI auth bypass less than 4 hours after disclosure. CVE-2026-44338 impacts versions 2.5.6 to 4.6.33; fix is in 4.6.34. #PraisonAI #CVE-2026-44338 #Sysdig https://ift.tt/GfmtBwT
Post summary
Sysdig observed scans of the PraisonAI authentication bypass (CVE-2026-44338) within hours of disclosure, and a fix is available in version 4.6.34.
00011134
4.3K followersView on X
西村/learningBOX/競プロｱｶ@ynishi2015149d
Active Exploitation
https://securityonline.info/praisonai-vulnerability-cve-2026-44338-exploited-in-the-wild/ セキュリティパッチがリリースされると、その差分を解析して数時間で攻撃が発生してしまうらしい。セキュリティパッチをリリースしないわけにはいかない。厳しい世界だ。
Post summary
CVE‑2026‑44338 has been reported as being actively exploited in the wild, prompting an urgent security patch release.
00020642
26.1K followersView on X
甲斐甲@k_aik_ou51d
Disclosure
📝 新着記事を公開しました認証オフで公開されたPraisonAIのAPIサーバー、3時間44分でスキャンされた GitHubで★8.3k・フォーク1.3kを集めるマルチエージェントフレームワーク PraisonAI に、認証バイパスの脆弱性（CVE-2026-44338）が公開されました。GitHub Security Advisory が公開されてから… https://t.co/kTuGuk1tUV
Post summary
The tweet announces the disclosure of an authentication bypass vulnerability (CVE-2026-44338) in the PraisonAI API server, citing a GitHub Security Advisory.
10000103
638 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	praison	praisonai	-	-	-

Explore more

Back to dashboard praison CVEs praisonai vulnerabilities

What researchers are saying

CVE-2026-44338

Sysdig[verified]@sysdigActive Exploitation

The message highlights that CVE-2026-44338 began being actively exploited only about 3.5 hours after disclosure, underscoring the rapid exploitation risk for AI-focused projects.

TodayInCyber[verified]@TodayInCyberIOActive Exploitation

Threat actors have exploited CVE-2026-44338 within hours of its public disclosure; no PoC, exploit code, patch, or false‑positive claim is provided, and the vulnerability is described as a critical authentication bypass.

connect24h[verified]@connect24hActive Exploitation

The report indicates that PraisonAI’s authentication bypass vulnerability (CVE‑2026‑44338, CVSS 7.3) was actively exploited within four hours of disclosure, underscoring the urgency of addressing AI infrastructure security.

DFIR Radar[verified]@DFIR_RadarActive Exploitation

CVE‑2026‑44338 was actively exploited within 4 hours, with an authentication bypass via hard‑coded settings in PraisonAI’s api_server.py, and WAF rules are recommended as a mitigation.

Blue Team News[verified]@blueteamsec1Active Exploitation

An authorization bypass vulnerability (CVE‑2026‑44338) in PraisonAI was reportedly targeted within hours of its public disclosure, indicating active exploitation in the wild.

Rock Lambros[verified]@rocklambrosGeneral

The tweet notes a vulnerability (CVE‑2026‑44338) was probed shortly after disclosure and cites a general exploitation statistic, without providing technical details, patches, or evidence of active exploitation.

Kwame[verified]@kwame_nyxActive Exploitation

CVE‑2026‑44338 was disclosed and quickly exploited, targeting the Flask API that had AUTH_ENABLED set to false, while no patch or exploit code is publicly referenced.

White Rabbitx 🏴‍☠️[verified]@TheRabbitPyPatch

A new authentication bypass vulnerability (CVE-2026-44338) affects PraisonAI, urging users to review deployments and patch promptly to prevent potential unauthorized access to sensitive AI agent workflows.