CVE-2026-25130Disclosure

LOWCVSS 9.6 · CRITICAL

Exploit discussion active in current signal (1 latest mentions)

Immediate actions

Hunt for exploitation attempts and persistence artifacts
Increase monitoring for publicly documented tradecraft
Track advisory updates for patch or workaround availability

Recommended action window: High priority (within 72h)

NVD description

Cybersecurity AI (CAI) is a framework for AI Security. In versions up to and including 0.5.10, the CAI (Cybersecurity AI) framework contains multiple argument injection vulnerabilities in its function tools. User-controlled input is passed directly to shell commands via `subprocess.Popen()` with `shell=True`, allowing attackers to execute arbitrary commands on the host system. The `find_file()` tool executes without requiring user approval because find is considered a "safe" pre-approved command. This means an attacker can achieve Remote Code Execution (RCE) by injecting malicious arguments (like -exec) into the args parameter, completely bypassing any human-in-the-loop safety mechanisms. Commit e22a1220f764e2d7cf9da6d6144926f53ca01cde contains a fix.

Back to dashboard

1.5/ 10 priority

Sources & remediation

Other references

Weakness type (CWE)

CWE-78

Priority

LOW

Exploitation

NONE

PoC

YES

Patch

NONE

Momentum

STABLE

Threat summary

Public PoC is present in monitored signal
7 mentions across 4 observed days
Momentum state: stable

What's happening

PoC mentioned or linked in 1 signal
Technical details provided in 6 signals
Disclosure: 5 classified signals
General: 1 classified signal
Peaked 3d ago at 4 mentions (2026-01-30); latest day: 1
7 total mentions across 4 days

Deep dive

Activity timeline7 mentions / 4d

Signal classification3 categories

Disclosure

571.4%

PoC

114.3%

General

114.3%

Referenced assets7 URLs

By indicator

Classification over time

Date	Total	Labels

2026-01-30	4	Disclosure4
2026-01-31	1	Disclosure1
2026-02-01	1	PoC1
2026-02-02	1	General1

Full discourse7 posts

Averlon@Averlon_ai248d
General
When a new vulnerability drops, chaos follows. Advisories flood in. Hot takes spread. Everyone asks: “How bad is this?” But the harder question is: "When does this vulnerability actually matter, and why?" Today, security teams spend 15–60 minutes per CVE jumping between: CVE reports, advisories, blog posts, vendor commentary, and research notes just to answer that. Vulnerability Intelligence exists to collapse that work. For each CVE, it breaks down: • what the vulnerability actually is • what an attacker would need to exploit it • whether user interaction or access is required • whether it’s being exploited in the wild This is the reasoning layer teams usually have to reconstruct manually. We’ve made this available as a free resource for security teams who need clarity before decisions. Live example: https://research.averlon.ai/vulnerability-intelligence/cve/CVE-2026-25130
Post summary
The post promotes a free resource that compiles and clarifies CVE information for security teams, but it does not mention any PoC, exploit code, patch, technical details, or evidence of active exploitation.
0003087
37 followersView on X
S.Komichevsen Matsuk@w4yh250d
PoC
CVE-2026-25130の解説がPoCというかthat's allな感が // CAI find_file Agent Tool has Command Injection Vulnerability Through Argument Injection · CVE-2026-25130 · GitHub Advisory Database https://github.com/advisories/GHSA-jfpc-wj3m-qw2m
Post summary
The post references CVE‑2026‑25130 as a command‑injection flaw, notes an accompanying explanation that resembles a proof‑of‑concept, but provides no exploit code, patch information, or evidence of active exploitation.
00000187
326 followersView on X
PulsePatch.io@pulsepatchio250d
Disclosure
CAI framework's find_file Agent Tool has a command injection vulnerability (CVE-2026-25130) through argument injection. Assess usage of this tool with untrusted input. #CAIFramework #InfoSec #Cybersecurity https://www.pulsepatch.io/posts/cve-2026-25130-cai-framework-command-injection
Post summary
The post discloses a command injection flaw (CVE‑2026‑25130) in CAI framework’s find_file agent tool and urges assessing its use with untrusted input.
0000073
1 followersView on X
CVEFind.com@CveFindCom251d
Disclosure
[CVE-2026-25130: CRITICAL] Beware of Cybersecurity AI (CAI) framework versions up to 0.5.10! Multiple argument injection vulnerabilities could allow attackers to execute commands on host system. Check commit...#cve,CVE-2026-25130,#cybersecurity https://cvefind.com/CVE-2026-25130
Post summary
The message alerts readers to a critical argument injection flaw in the Cybersecurity AI framework (v0.5.10 and earlier) that could enable host command execution, but offers no PoC, exploit, or patch information.
0000051
584 followersView on X
The Hacker Wire@TheHackerWire251d
Disclosure
🔴 CVE-2026-25130 - Critical Cybersecurity AI (CAI) is a framework for AI Security. In versions up to and including 0.5.10, the CAI (Cybersecurity AI) framework contains multiple argument injection vulnerabilities in... https://www.thehackerwire.com/vulnerability/CVE-2026-25130/ https://t.co/0hOw2wf62o
Post summary
The tweet announces CVE‑2026‑25130 as a critical argument injection flaw affecting CAI framework versions up to 0.5.10, with a link to further details.
0000060
113 followersView on X
CVE@CVEnew251d
Disclosure
CVE-2026-25130 Cybersecurity AI (CAI) is a framework for AI Security. In versions up to and including 0.5.10, the CAI (Cybersecurity AI) framework contains multiple argument injecti… https://www.cve.org/CVERecord?id=CVE-2026-25130
Post summary
The passage discloses CVE-2026-25130 as a vulnerability in Cybersecurity AI (CAI) involving argument injection in versions up to 0.5.10, without providing any PoC, exploit, or mitigation details.
00000174
56.5K followersView on X
0day Signal@0dayPublishing251d
Disclosure
🚨 CVE-2026-25130: Cybersecurity AI vulnerable to c... Classic subprocess.Popen() with shell=True strikes again – CAI's "safe" find_file() tool enables unapproved RCE via arg... https://zerodaysignal.com/vulnerability/CVE-2026-25130 #netsec #vulnerability #CVE #sysadmin #zeroday
Post summary
The tweet announces CVE‑2026‑25130, detailing an RCE via the CAI find_file() tool using subprocess.Popen(), but lacks a patch notice, PoC, or evidence of active exploitation.
0000073
132 followersView on X

Explore more

Back to dashboard

What researchers are saying

CVE-2026-25130

Averlon[verified]@Averlon_aiGeneral

The post promotes a free resource that compiles and clarifies CVE information for security teams, but it does not mention any PoC, exploit code, patch, technical details, or evidence of active exploitation.

S.Komichevsen Matsuk@w4yhPoC

The post references CVE‑2026‑25130 as a command‑injection flaw, notes an accompanying explanation that resembles a proof‑of‑concept, but provides no exploit code, patch information, or evidence of active exploitation.

PulsePatch.io@pulsepatchioDisclosure

The post discloses a command injection flaw (CVE‑2026‑25130) in CAI framework’s find_file agent tool and urges assessing its use with untrusted input.

CVEFind.com@CveFindComDisclosure

The message alerts readers to a critical argument injection flaw in the Cybersecurity AI framework (v0.5.10 and earlier) that could enable host command execution, but offers no PoC, exploit, or patch information.

The Hacker Wire@TheHackerWireDisclosure

The tweet announces CVE‑2026‑25130 as a critical argument injection flaw affecting CAI framework versions up to 0.5.10, with a link to further details.

CVE@CVEnewDisclosure

The passage discloses CVE-2026-25130 as a vulnerability in Cybersecurity AI (CAI) involving argument injection in versions up to 0.5.10, without providing any PoC, exploit, or mitigation details.

0day Signal@0dayPublishingDisclosure

The tweet announces CVE‑2026‑25130, detailing an RCE via the CAI find_file() tool using subprocess.Popen(), but lacks a patch notice, PoC, or evidence of active exploitation.