Exploit catalog results

Showing 19 PoCs on this page

ExploitDB

PraisonAI praisonaiagents 1.6.77 - Remote Code Execution

ExploitDB exploitPublished 2026-08-11
ExploitCVE-2026-614471 file

EDB-52639

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A Python script that directly instantiates a vulnerable CodeAgent and passes it malicious code to exfiltrate environment secrets, demonstrating remote code execution.

Backdoor review

No backdoor observed in reviewed code

The PoC script demonstrates the claimed CVE-2026-61447 vulnerability by importing the vulnerable CodeAgent and executing a hardcoded payload that reads environment variables. The payload is limited to printing secrets to stdout; no network exfiltration, persistence, or unrelated harmful behavior is present. The script is a straightforward, transparent exploit demonstration with no concealed or deceptive functionality.

ClassificationExploit
Model confidence95%
AuthenticationNot required
Languagespython
Target softwarePraisonAI praisonaiagents
Attack typesremote code executioninformation disclosure
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script's primary purpose is to exercise the CVE-2026-61447 vulnerability by directly executing attacker-controlled code on the target system. It imports the vulnerable component, crafts a payload to exfiltrate secrets, and executes it, which constitutes an exploit.

exploits/multiple/remote/52639.py:2exploits/multiple/remote/52639.py:18exploits/multiple/remote/52639.py:54-70

Requirements

  • Requires the vulnerable praisonaiagents package (<= 1.6.77) installed.exploits/multiple/remote/52639.py:30-31

Observed behavior

  • Imports and instantiates the vulnerable CodeAgent class.exploits/multiple/remote/52639.py:54-57
  • Defines malicious Python code that imports os and json, then filters and prints environment variables containing keywords like 'KEY', 'SECRET', 'TOKEN', etc.exploits/multiple/remote/52639.py:59-67
  • Executes the malicious code by calling agent.execute(malicious_code).exploits/multiple/remote/52639.py:70
  • Prints the execution result, including stdout, stderr, and return code.exploits/multiple/remote/52639.py:72-75
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Environment Variable Access
Payload withheldThe hardcoded payload reads environment variables containing keywords like KEY, SECRET, TOKEN, PASSWORD, CREDENTIAL, API, AUTH and prints them to stdout. This is the claimed vulnerability demonstration and does not exfiltrate data externally.exploits/multiple/remote/52639.py:60-66
Import
Payload withheldThe script imports the vulnerable component to trigger the RCE. This is expected for a PoC targeting the stated CVE.exploits/multiple/remote/52639.py:54
Review boundaries

What the analysis did not establish

  • Analysis is based solely on the static content of the provided exploit file; the code was not executed.
  • The exploit's effectiveness depends on the presence of the vulnerable library version and a functional environment, which cannot be verified from the file alone.
  • Only the supplied exploit script text was reviewed; the imported praisonaiagents library and its CodeAgent implementation were not inspected.
  • Binary or non-text content was not present in the evidence.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

mcp-server-kubernetes 3.8.x - Argument Injection

ExploitDB exploitPublished 2026-08-11
ExploitCVE-2026-614591 file

EDB-52638

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that sends a crafted JSON-RPC request to an MCP Server Kubernetes endpoint, injecting the --server flag via the resourceType parameter to redirect kubectl commands to an attacker-controlled API server, thereby exfiltrating the operator's bearer token.

Backdoor review

No backdoor observed in reviewed code

The PoC script sends a crafted JSON-RPC request to exploit CVE-2026-61459 by injecting a --server flag into the resourceType parameter. It requires the operator to supply both a target MCP server and an attacker-controlled server URL. The script only makes a single HTTP POST to the target; it does not connect to the attacker server, exfiltrate data, establish persistence, or execute any unrelated payload. No concealed or deceptive behavior was found.

ClassificationExploit
Model confidence95%
AuthenticationNot required
Languagespython
Target softwaremcp-server-kubernetes
Attack typesargument injectioncredential exfiltration
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively sends a payload designed to inject arguments into a server-side command, causing the server to connect to an attacker-controlled host and leak credentials. This constitutes exploitation, not mere detection or scanning.

exploits/multiple/remote/52638.py:68-82exploits/multiple/remote/52638.py:88-100

Requirements

  • Requires network access to the MCP Server JSON-RPC endpoint.exploits/multiple/remote/52638.py:62
  • Requires an attacker-controlled Kubernetes API server to receive the exfiltrated token.exploits/multiple/remote/52638.py:63

Observed behavior

  • Constructs a JSON-RPC tools/call request with a malicious resourceType value containing --server=<attacker_url>.exploits/multiple/remote/52638.py:69-82
  • Sends the crafted request to the target MCP server via HTTP POST.exploits/multiple/remote/52638.py:89-94
  • Reports success based on HTTP response status and instructs the user to check a listener for the exfiltrated bearer token.exploits/multiple/remote/52638.py:96-100
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Network Connection
Payload withheldThe script sends the exploit payload to the target MCP server. It does not connect to the attacker server.exploits/multiple/remote/52638.py:89-94
Payload
Payload withheldThis is the argument injection payload that exploits the vulnerability. It is the core of the PoC and is expected behavior.exploits/multiple/remote/52638.py:75
Review boundaries

What the analysis did not establish

  • Analysis is based solely on the static content of the provided Python script; no external dependencies (e.g., requests library) were resolved, and the code was not executed.
  • The script's effectiveness depends on the target server's configuration and the presence of a listener on the attacker's machine, which are not verified by the evidence.
  • Only the supplied text file was reviewed; no external dependencies (e.g., the requests library) were inspected.
  • The review does not assess whether the target MCP server is actually vulnerable or whether the exploit works as claimed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Blocksy Companion 2.1.46 - RCE

ExploitDB exploitPublished 2026-08-11
ExploitCVE-2026-584801 file

EDB-52640

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that sends an unauthenticated POST request to upload a PHP web shell (poc.woff2.php) via the blc-review-images[] parameter to the admin-ajax.php endpoint, exploiting a double-extension file upload bypass in Blocksy Companion <= 2.1.46.

Backdoor review

No backdoor observed in reviewed code

The PoC is a straightforward unauthenticated file upload exploit for CVE-2026-58480. It uploads a minimal PHP webshell that executes commands passed via a GET parameter. No concealed behavior, persistence mechanisms, credential theft, or unrelated payloads are present. The code is fully readable and contains no obfuscation or deceptive functionality.

ClassificationExploit
Model confidence100%
AuthenticationNot required
Languagespython
Target softwareBlocksy CompanionWordPress
Attack typesremote code executionunrestricted file upload
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively uploads a PHP web shell to the target server and provides instructions for executing arbitrary commands, which constitutes exploitation rather than mere detection or scanning.

exploits/multiple/webapps/52640.py:61-69exploits/multiple/webapps/52640.py:79-85

Requirements

  • Target must run Blocksy Companion <= 2.1.46 with Advanced Reviews and Custom Fonts extensions enabled.exploits/multiple/webapps/52640.py:12exploits/multiple/webapps/52640.py:28
  • Target URL must be supplied as a command-line argument.exploits/multiple/webapps/52640.py:53-56

Observed behavior

  • Constructs a PHP payload that executes system commands via the 'cmd' GET parameter.exploits/multiple/webapps/52640.py:62-69
  • Sends an unauthenticated POST request to /wp-admin/admin-ajax.php with the payload as blc-review-images[] and action=blc_save_review_attachments.exploits/multiple/webapps/52640.py:71-81
  • Prints the HTTP status and response, then instructs the user to access the uploaded shell in wp-content/uploads/.exploits/multiple/webapps/52640.py:82-85
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Webshell Payload
Payload withheldThe uploaded payload is a simple PHP webshell that executes commands supplied via the 'cmd' GET parameter. This is the intended exploit behavior for the stated CVE and does not constitute a backdoor beyond the advertised RCE.exploits/multiple/webapps/52640.py:62-69
Exploit Mechanism
Payload withheldThe exploit uses the documented vulnerability to upload a file with a double extension (poc.woff2.php) to bypass extension validation. This is consistent with the CVE description and does not involve any hidden or unrelated actions.exploits/multiple/webapps/52640.py:71-81
Review boundaries

What the analysis did not establish

  • Only the exploit script is provided; external dependencies (requests library) and the target environment are not included.
  • The script's actual effectiveness is not verified; classification is based solely on the code's apparent intent.
  • Only the exploit script text was reviewed; no network traffic or runtime behavior was observed.
  • The review does not assess the safety or reliability of the exploit, only the absence of backdoor behavior within the provided code.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

LuCI DHCPv6 - Lease Hostname Stored Cross-Site Scripting

ExploitDB exploitPublished 2026-08-11
ExploitCVE-2026-618761 file

EDB-52637

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that sends crafted DHCPv6 messages containing a malicious hostname in the Client FQDN option (option 39) to inject a stored XSS payload into the LuCI web interface of OpenWrt.

Backdoor review

No backdoor observed in reviewed code

The supplied Python script is a proof-of-concept for CVE-2026-61876, a stored XSS vulnerability in LuCI DHCPv6 lease hostname handling. It constructs and sends DHCPv6 messages with a user-supplied hostname payload to a target server. The code performs only the advertised DHCPv6 interaction; no backdoor, persistence, credential theft, unrelated remote access, or concealed harmful behavior was observed.

ClassificationExploit
Model confidence98%
AuthenticationNot required
Languagespython
Target softwareOpenWrt LuCIodhcpd
Attack typesstored cross-site scripting
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively constructs and transmits malicious DHCPv6 network packets to inject a stored XSS payload into a target system. This is code intended to exercise a vulnerability, not merely detect it.

exploits/multiple/dos/52637.py:1exploits/multiple/dos/52637.py:15exploits/multiple/dos/52637.py:17exploits/multiple/dos/52637.py:23exploits/multiple/dos/52637.py:183-189

Requirements

  • Adjacent network access (LAN) to the target OpenWrt device.exploits/multiple/dos/52637.py:32
  • An administrator must view the DHCP lease status page in LuCI to trigger the payload.exploits/multiple/dos/52637.py:33

Observed behavior

  • Constructs and sends a DHCPv6 Solicit message containing a user-supplied hostname in the Client FQDN option.exploits/multiple/dos/52637.py:183-189
  • Processes the DHCPv6 Advertise response to extract the server ID and IA_NA options.exploits/multiple/dos/52637.py:190-198
  • Sends a DHCPv6 Request message to confirm the lease, again including the malicious FQDN.exploits/multiple/dos/52637.py:203-211
  • Optionally sends a DHCPv6 Release message to release the acquired lease.exploits/multiple/dos/52637.py:220-228
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Network Behavior
Payload withheldThis is the expected behavior for a DHCPv6 client PoC exploiting a hostname injection vulnerability.exploits/multiple/dos/52637.py:183-189exploits/multiple/dos/52637.py:203-211exploits/multiple/dos/52637.py:220-228
Payload Injection
Payload withheldThis directly implements the described XSS attack vector by injecting the payload into the DHCPv6 lease hostname field.exploits/multiple/dos/52637.py:176exploits/multiple/dos/52637.py:188exploits/multiple/dos/52637.py:210
Review boundaries

What the analysis did not establish

  • Only the single Python file from the ExploitDB entry was analyzed; no external libraries or referenced resources were expanded.
  • The code was not executed; classification is based solely on static analysis of the source code and its stated purpose.
  • Only the single Python file was reviewed; no external dependencies or runtime environment were inspected.
  • The script's effect depends on the target server's behavior, which is outside the scope of this review.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

OrkesConductor 3.30.2 - Unauthenticated Remote Code Execution

ExploitDB exploitPublished 2026-08-10
ExploitCVE-2026-581381 file

EDB-52633

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that exploits CVE-2026-58138 in Orkes Conductor to achieve unauthenticated remote code execution by submitting a malicious INLINE JavaScript task that uses Java reflection to execute arbitrary OS commands.

Backdoor review

No backdoor observed in reviewed code

The PoC is a straightforward Python exploit for CVE-2026-58138. It constructs a malicious JavaScript payload that uses Java reflection to execute a user-supplied OS command on the target Conductor server, then retrieves the output. All behavior is consistent with the stated purpose of demonstrating unauthenticated RCE. No concealed executable behavior, persistence, credential theft, or unrelated payloads were observed.

ClassificationExploit
Model confidence98%
AuthenticationNot required
Languagespython
Target softwareOrkes ConductorConductor OSS
Attack typesremote code executioncode injection
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script is a complete, self-contained Python program that actively exploits the vulnerability by sending crafted HTTP requests to execute arbitrary OS commands on the target. It does not merely detect or report the vulnerability; it performs the exploitation and retrieves command output.

exploits/multiple/webapps/52633.py:1-139

Requirements

  • Target running vulnerable Orkes Conductor (3.21.21 to <3.30.2) with default community API behavior (no authentication required).exploits/multiple/webapps/52633.py:12exploits/multiple/webapps/52633.py:32

Observed behavior

  • Constructs a JavaScript expression that uses Java reflection to obtain java.lang.Runtime, build a String array containing 'sh', '-c', and the attacker-supplied command, then invokes Runtime.exec() to execute the command.exploits/multiple/webapps/52633.py:55-73
  • Registers a new workflow definition containing the malicious INLINE task via an unauthenticated POST to /api/metadata/workflow.exploits/multiple/webapps/52633.py:100-117
  • Starts the workflow via POST to /api/workflow/<name>, then fetches the workflow output to retrieve the command execution result.exploits/multiple/webapps/52633.py:118-133
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Command Execution
Payload withheldThis is the core exploit mechanism for the claimed vulnerability.exploits/multiple/webapps/52633.py:55-73
Network Request
Payload withheldThis is the delivery and retrieval mechanism for the exploit.exploits/multiple/webapps/52633.py:76-90exploits/multiple/webapps/52633.py:117-118
Review boundaries

What the analysis did not establish

  • Evidence is limited to the single Python script; no external dependencies, network captures, or execution logs are included.
  • The script's effectiveness depends on the target's configuration and the presence of the vulnerability; no verification of successful exploitation is provided in the evidence.
  • Only the single Python file was reviewed; no external dependencies or referenced resources were inspected.
  • The review does not assess the safety or correctness of the exploit against the target system.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Microsoft Edge 150.0.4078.48 - RCE

ExploitDB exploitPublished 2026-08-10
ScannerCVE-2026-582891 file

EDB-52632

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a Python HTTP server that serves a static HTML page containing JavaScript. The JavaScript attempts to trigger a type confusion vulnerability in the V8 engine by repeatedly calling a function with an object and then assigning a value to a property, but it does not contain any shellcode, memory corruption primitives, or code execution payload. The author's notes explicitly state it is a 'Proof of Concept only' and 'No public full exploit is available yet.' The code's primary operation is to serve a test page that checks for a crash or observable behavior, consistent with a vulnerability scanner or trigger, not a working exploit.

Backdoor review

No backdoor observed in reviewed code

The PoC is a simple Python HTTP server that delivers a static HTML page containing JavaScript intended to trigger a type confusion vulnerability in Microsoft Edge. The code performs no file system writes, network connections beyond the local HTTP server, persistence mechanisms, or delivery of unrelated payloads. No obfuscation or concealed behavior is present.

ClassificationScanner
Model confidence95%
AuthenticationNot required
LanguagesPythonJavaScriptHTML
Target softwareMicrosoft Edge (Chromium-based)
Attack typesType Confusion
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The artifact is classified as a scanner because its primary operation is to serve a test page that attempts to trigger a type confusion condition and instructs the user to monitor for crashes, without containing any exploit payload, memory manipulation, or code execution logic. The author explicitly states it is a 'Proof of Concept only' and that 'No public full exploit is available yet.'

exploits/multiple/local/52632.py`:29exploits/multiple/local/52632.py`:42exploits/multiple/local/52632.py`:70-78

Requirements

  • Victim must visit the attacker-controlled HTTP server with a vulnerable version of Microsoft Edge (before 150.0.4078.48).exploits/multiple/local/52632.py`:11exploits/multiple/local/52632.py`:39

Observed behavior

  • Starts an HTTP server on port 8080 (or a custom port) and serves a static HTML page at '/' or '/poc.html'.exploits/multiple/local/52632.py`:95-101exploits/multiple/local/52632.py`:106-118
  • The served HTML page contains JavaScript that creates an object, calls a dummy function on it 10,000 times, then assigns the value 0x41414141 to a property of the object, logging a message to the console.exploits/multiple/local/52632.py`:70-78
  • The author's notes instruct the user to monitor the browser process for crashes or code execution, indicating the code is intended to test for a vulnerability rather than reliably achieve code execution.exploits/multiple/local/52632.py`:29exploits/multiple/local/52632.py`:42
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Network Service
Payload withheldThe PoC starts a local HTTP server to deliver the exploit page to a vulnerable browser.exploits/multiple/local/52632.py`:106-118
Javascript Payload
Payload withheldThe JavaScript attempts to trigger CVE-2026-58289 by manipulating object types in a loop and assigning a value to a property.exploits/multiple/local/52632.py`:70-78
Review boundaries

What the analysis did not establish

  • The analysis is based solely on the static source code; the code was not executed, and its actual behavior in a vulnerable environment is not verified.
  • The JavaScript payload is simplistic and lacks any demonstrated memory corruption or code execution capability, consistent with a trigger or scanner rather than a functional exploit.
  • Only the supplied text file was reviewed; no external resources, libraries, or binary files were inspected.
  • The JavaScript payload is a simplified trigger and may not reliably exploit the vulnerability; its effectiveness is not assessed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Flowise 3.1.3 - arbitrary code execution

ExploitDB exploitPublished 2026-07-07
ScannerCVE-2026-580571 file

EDB-52623

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The script validates the CVE-2026-58057 case-sensitivity bypass locally without interacting with a Flowise instance. It simulates the vulnerable and fixed validation logic, then uses a local Node.js canary to confirm that the lower-case 'payload withheld' environment variable is honored on Windows, demonstrating the bypass condition.

Backdoor review

No backdoor observed in reviewed code

The PoC script demonstrates the CVE-2026-58057 vulnerability by validating the case-sensitivity bypass and performing a local canary test. It writes a temporary loader file, sets the 'payload withheld' environment variable, and runs Node.js to check if the bypass works. No concealed executable behavior, persistence, credential theft, or unrelated remote access is observed. The script's actions are consistent with a proof-of-concept for the described vulnerability.

ClassificationScanner
Model confidence95%
AuthenticationNot required
Languagespython
Target softwareFlowise
Attack typesremote code execution
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script does not connect to or exploit a Flowise server. It locally validates the vulnerability condition by simulating the denylist logic and testing the Node.js runtime behavior. This is detection/validation, not exploitation.

exploits/multiple/webapps/52623.py:1-157

Requirements

  • Node.js must be installed and available in PATH for the canary test.exploits/multiple/webapps/52623.py:64-66
  • The script must be executed on a Windows system to reproduce the case-insensitivity bypass.exploits/multiple/webapps/52623.py:127

Observed behavior

  • Simulates the vulnerable Flowise validation function, which uses case-sensitive set membership to block dangerous environment variables.exploits/multiple/webapps/52623.py:47-53
  • Simulates a fixed validation function that normalizes keys to uppercase before checking the denylist.exploits/multiple/webapps/52623.py:55-61
  • Tests whether the lower-case 'payload withheld' bypasses the vulnerable validator and whether the normalized validator blocks it.exploits/multiple/webapps/52623.py:107-121
  • Creates a temporary Node.js loader script and spawns a Node.js process with 'payload withheld' set to --require that loader. The loader writes a canary file if the environment variable is honored.exploits/multiple/webapps/52623.py:63-100
  • Outputs a JSON result indicating whether the bypass was reproduced, including the canary file creation status.exploits/multiple/webapps/52623.py:125-136
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

File Write
Payload withheldThe script writes a temporary JavaScript loader file to disk to test the payload withheld bypass.exploits/multiple/webapps/52623.py:70-73
Environment Variable Set
Payload withheldThe script sets the 'payload withheld' environment variable to demonstrate the case-sensitivity bypass.exploits/multiple/webapps/52623.py:78
Process Execution
Payload withheldThe script executes Node.js with the manipulated environment to test if the bypass works.exploits/multiple/webapps/52623.py:84-90
Review boundaries

What the analysis did not establish

  • The script does not interact with a Flowise instance; it only validates the bypass condition locally.
  • The canary test requires Node.js to be installed and the script to run on Windows; on other platforms the finding_reproduced flag may be false even if the validation bypass is present.
  • The script does not demonstrate actual code execution within a Flowise server context.
  • Only the supplied text file was reviewed; no external dependencies or binary files were inspected.
  • The script's behavior is limited to local testing and does not include the actual exploitation step against a remote Flowise instance.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Hydra - Stack Buffer Overflow

ExploitDB exploitPublished 2026-07-07
ExploitCVE-2026-567661 file

EDB-52622

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that implements a malicious SMTP server to trigger a stack buffer overflow in Hydra's NTLM authentication handler by sending a crafted NTLM Type-2 challenge with an excessively long domain name.

Backdoor review

No backdoor observed in reviewed code

The PoC is a straightforward SMTP server that sends a crafted NTLM Type-2 challenge to trigger a known stack buffer overflow in vulnerable Hydra clients. No backdoor, deceptive payload, or concealed operator-directed harm is present.

ClassificationExploit
Model confidence95%
AuthenticationNot required
Languagespython
Target softwareHydra
Attack typesstack buffer overflow
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively sends a malicious payload (crafted NTLM Type-2 challenge) designed to overflow a stack buffer in a vulnerable client, which is the definition of an exploit. It does not merely check for the vulnerability; it exercises it to cause a crash (SIGSEGV).

exploits/linux/remote/52622.py:1exploits/linux/remote/52622.py:15exploits/linux/remote/52622.py:17-19exploits/linux/remote/52622.py:29exploits/linux/remote/52622.py:57-76exploits/linux/remote/52622.py:100-103

Requirements

  • A vulnerable Hydra client (<= 9.7) must connect to the malicious server and attempt NTLM authentication.exploits/linux/remote/52622.py:11exploits/linux/remote/52622.py:28exploits/linux/remote/52622.py:33-34

Observed behavior

  • Starts a TCP server on port 2525 that speaks SMTP and advertises NTLM authentication.exploits/linux/remote/52622.py:78-83exploits/linux/remote/52622.py:95
  • When a client sends an NTLM authentication token, the server responds with a crafted NTLM Type-2 challenge containing a 400-byte domain name.exploits/linux/remote/52622.py:57-76exploits/linux/remote/52622.py:100-103
  • The oversized domain name causes a stack buffer overflow in the Hydra client when it processes the base64-encoded challenge, leading to a SIGSEGV.exploits/linux/remote/52622.py:17-19exploits/linux/remote/52622.py:29
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Network Server
Payload withheldThe PoC acts as a malicious SMTP server to deliver the overflow payload.exploits/linux/remote/52622.py:78-83
Exploit Trigger
Payload withheldThe overflow is triggered by a long domain string in the NTLM challenge, matching the CVE description.exploits/linux/remote/52622.py:57-76
Review boundaries

What the analysis did not establish

  • The analysis is based solely on the provided Python script; no external dependencies, libraries, or the vulnerable Hydra binary were inspected or executed.
  • The script's effectiveness and reliability are not verified; it is treated as untrusted data.
  • Only the supplied text file was reviewed; no binary or external resources were inspected.
  • The PoC does not include a payload for remote code execution; it only triggers a crash (SIGSEGV) as stated in the comments.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Discuz! X5.0 - Authentication Bypass

ExploitDB exploitPublished 2026-07-07
ExploitCVE-2026-499521 file

EDB-52621

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that exploits CVE-2026-49952 in Discuz! X5.0 by sending a crafted login request to obtain an authcode via an encryption oracle, then using that authcode to access the dbbak.php database backup API without authentication.

Backdoor review

No backdoor observed in reviewed code

The PoC script performs only the documented authentication bypass exploit against Discuz! X5.0. It sends a crafted login request to obtain an authcode and then uses that authcode to access the dbbak.php backup API. No backdoor, persistence, credential exfiltration, or unrelated payload behavior is present.

ClassificationExploit
Model confidence95%
AuthenticationNot required
Languagespython
Target softwareDiscuz! X5.0
Attack typesauthentication bypassencryption oracle abuse
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively performs the steps to exploit the vulnerability: it crafts a malicious login request to obtain a valid token (authcode) and then uses that token to access a protected database backup endpoint. This constitutes exploit code, not merely detection or scanning.

exploits/multiple/webapps/52621.py:60-92

Requirements

  • Target running vulnerable Discuz! X5.0 (versions 20260320 through 20260501) with dbbak.php exposed.exploits/multiple/webapps/52621.py:11

Observed behavior

  • Sends a POST request to member.php login endpoint with a crafted username payload (default 'admin|1|0|0') to trigger the encryption oracle and extract an authcode from the response.exploits/multiple/webapps/52621.py:60-78
  • Uses the extracted authcode to send a GET request to /api/db/dbbak.php with operation=backup, bypassing authentication to access database backup functionality.exploits/multiple/webapps/52621.py:80-92
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Network Request
Payload withheldExploit step to obtain authcode via encryption oracleexploits/multiple/webapps/52621.py:61-69
Network Request
Payload withheldExploit step to access database backup functionalityexploits/multiple/webapps/52621.py:81-84
Review boundaries

What the analysis did not establish

  • Only the exploit script content is analyzed; external dependencies (requests, urllib3) and target behavior are not verified.
  • The script's effectiveness depends on the target's specific configuration and the presence of the vulnerable dbbak.php endpoint.
  • Only the supplied text file was reviewed; no binary or external resources were inspected.
  • The script's behavior depends on the target server's response; no server-side backdoor is claimed or observed in the client code.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Tenable Nessus 10.12.1 - SQL Injection

ExploitDB exploitPublished 2026-07-07
ExploitCVE-2026-575881 file

EDB-52620

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that generates a malicious .nessus XML file containing SQL injection payloads in host property tags and plugin output. The file is intended to be imported by a privileged Nessus user to trigger SQL injection in the backend database.

Backdoor review

No backdoor observed in reviewed code

The PoC script generates a malicious .nessus XML file containing SQL injection payloads designed to exploit CVE-2026-57588. The script itself performs no backdoor, persistence, credential theft, or unrelated remote access actions. It only creates a file with the described exploit payloads and prints status messages.

ClassificationExploit
Model confidence95%
AuthenticationNot required
Languagespython
Target softwareTenable Nessus
Attack typessql injection
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script constructs a malicious input file containing SQL injection payloads designed to be processed by a vulnerable Nessus instance, which constitutes exploit code intended to exercise the vulnerability.

exploits/multiple/webapps/52620.py:1exploits/multiple/webapps/52620.py:17-18exploits/multiple/webapps/52620.py:64-68exploits/multiple/webapps/52620.py:81-88

Requirements

  • Requires social engineering to trick a privileged user into importing the generated .nessus file via the Nessus web interface.exploits/multiple/webapps/52620.py:31exploits/multiple/webapps/52620.py:39-40

Observed behavior

  • Generates a .nessus XML file with SQL injection payloads embedded in host property tag values (e.g., UNION SELECT, pg_sleep, pg_read_file) and in plugin_output text.exploits/multiple/webapps/52620.py:64-68exploits/multiple/webapps/52620.py:81-88
  • Writes the crafted XML to a user-specified output file.exploits/multiple/webapps/52620.py:90-91
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Sql Injection Payload
Payload withheldPayload embedded in the generated .nessus file to exfiltrate database information when imported by a victim.exploits/multiple/webapps/52620.py:64
Sql Injection Payload
Payload withheldTime-based blind SQL injection payload in the generated file.exploits/multiple/webapps/52620.py:65
Sql Injection Payload
Payload withheldBoolean-based SQL injection payload in the generated file.exploits/multiple/webapps/52620.py:66
Sql Injection Payload
Payload withheldPayload attempting to read a local file on the database server via SQL injection.exploits/multiple/webapps/52620.py:67
Sql Injection Payload
Payload withheldMulti-line SQL injection payload in the generated file designed to enumerate database schema.exploits/multiple/webapps/52620.py:81-88
Review boundaries

What the analysis did not establish

  • Only the exploit script is provided; the vulnerable Nessus application and database backend are not included.
  • The script generates a payload file but does not perform the import or demonstrate successful exploitation.
  • Only the provided text of the PoC script was reviewed; no external dependencies, libraries, or the generated .nessus file were inspected.
  • The review does not assess the safety or functionality of the generated payloads when imported into Nessus.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Repetier-Server 1.4.10 - Path Traversal

ExploitDB exploitPublished 2026-04-30
ScannerCVE-2026-263351 file

EDB-52540

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a Python script that attempts to read arbitrary files from a Repetier-Server target via path traversal. It sends HTTP GET requests with traversal payloads and checks the response status and size to determine success, but does not deliver a payload or execute code on the target.

Backdoor review

No backdoor observed in reviewed code

The PoC is a straightforward path traversal exploit script for Repetier-Server. It constructs traversal payloads, sends HTTP GET requests, and prints the response. No backdoor, deceptive payload, persistence, credential theft, or unrelated remote access behavior is present. The script includes a standard warning to test only on authorized systems.

ClassificationScanner
Model confidence95%
AuthenticationNot required
LanguagesPython
Target softwareRepetier-Server
Attack typesPath Traversal
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script only sends HTTP requests to read files and reports success based on response characteristics. It does not deliver a payload, establish a shell, or execute commands on the target. This is consistent with a scanner that checks for the vulnerability without exploiting it.

exploits/multiple/webapps/52540.py:20-50

Requirements

  • Target must be running Repetier-Server version <= 1.4.10exploits/multiple/webapps/52540.py:1
  • Target URL must be reachable over HTTPexploits/multiple/webapps/52540.py:37

Observed behavior

  • Sends HTTP GET requests with path traversal sequences to read files from the target serverexploits/multiple/webapps/52540.py:23-26exploits/multiple/webapps/52540.py:37
  • Checks response status code 200 and content length > 60 bytes to determine if file read was successfulexploits/multiple/webapps/52540.py:39
  • Prints a preview of the retrieved file content if successfulexploits/multiple/webapps/52540.py:40-42
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Url
Payload withheldAdvisory URL referenced in a comment; not fetched or executed by the script.exploits/multiple/webapps/52540.py:7
Review boundaries

What the analysis did not establish

  • The CVE context (CVE-2026-26335) describes a static machineKey vulnerability in Calero VeraSMART, but the PoC targets Repetier-Server path traversal, indicating a possible CVE mismatch in the metadata.
  • The script uses 'verify=False' in requests.get, which disables SSL certificate verification and may mask connection issues, but this does not affect classification.
  • Only the supplied text file was reviewed; no binary files or external resources were inspected.
  • The script uses 'verify=False' in requests.get, disabling TLS certificate validation, which is a security weakness but not a backdoor.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

deephas 1.0.7 - Prototype Pollution

ExploitDB exploitPublished 2026-04-30
ExploitCVE-2026-250471 file

EDB-52528

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that executes two Node.js payloads to pollute Object.prototype via deephas.set() using constructor.prototype and __proto__ paths, confirming vulnerability by reading polluted properties on empty objects.

Backdoor review

No backdoor observed in reviewed code

The PoC script is a straightforward demonstration of a prototype pollution vulnerability in deephas 1.0.7. It constructs two JavaScript payloads that pollute Object.prototype and then checks for the pollution. The Python wrapper only writes the payloads to a temp file, executes them with Node.js, captures output, and cleans up. No backdoor, persistence, credential theft, unrelated remote access, or concealed harmful behavior is present.

ClassificationExploit
Model confidence95%
AuthenticationNot required
LanguagesPythonJavaScript
Target softwaredeephas
Attack typesPrototype Pollution
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively exercises the prototype pollution vulnerability by calling deephas.set() with malicious paths and verifying the resulting pollution on Object.prototype, which constitutes exploitation rather than mere detection.

exploits/multiple/webapps/52528.py:121-131exploits/multiple/webapps/52528.py:137-147

Requirements

  • Node.js must be installed and available in PATHexploits/multiple/webapps/52528.py:113-115
  • Vulnerable deephas@1.0.7 must be installed via npmexploits/multiple/webapps/52528.py:117-118

Observed behavior

  • Overwrites Object.prototype.hasOwnProperty to always return true, then calls deephas.set() with 'constructor.prototype.poc1' path and checks if {}.poc1 equals 'yes!!!'exploits/multiple/webapps/52528.py:121-131
  • Overwrites String.prototype.indexOf to always return -1, then calls deephas.set() with '__proto__.poc2' path and checks if {}.poc2 equals 'HACKED'exploits/multiple/webapps/52528.py:137-147
  • Prints VULNERABLE status if polluted values are observed on empty objectsexploits/multiple/webapps/52528.py:96-98
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

File Write
Payload withheldThe script writes JavaScript code to a temporary file for execution.exploits/multiple/webapps/52528.py:53-57
Process Execution
Payload withheldThe script executes the temporary JavaScript file using Node.js.exploits/multiple/webapps/52528.py:59-65
File Deletion
Payload withheldThe script deletes the temporary file after execution.exploits/multiple/webapps/52528.py:80-84
Prototype Pollution Payload
Payload withheldThe first PoC payload demonstrates prototype pollution via constructor.prototype.exploits/multiple/webapps/52528.py:122-126
Prototype Pollution Payload
Payload withheldThe second PoC payload demonstrates prototype pollution via __proto__.exploits/multiple/webapps/52528.py:138-142
Review boundaries

What the analysis did not establish

  • Only the Python wrapper and embedded JavaScript payloads were analyzed; the deephas library source code was not inspected.
  • The script's actual execution and pollution success were not verified; analysis is based solely on static code review.
  • The review is limited to the supplied Python script. The deephas npm package and the Node.js runtime are not inspected. The script's behavior depends on the installed version of deephas, which is not verified by the script itself.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

JUNG Smart Visu Server 1.1.1050 - Dos

ExploitDB exploitPublished 2026-04-30
ExploitCVE-2026-262351 file

EDB-52536

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that sends an unauthenticated payload withheld request to payload withheld or payload withheld on a JUNG Smart Visu Server to trigger a remote reboot or shutdown, causing denial of service.

Backdoor review

No backdoor observed in reviewed code

The PoC script sends an unauthenticated payload withheld request to payload withheld or payload withheld on the target JUNG Smart Visu Server to trigger a denial-of-service condition. No backdoor, deceptive payload, persistence mechanism, credential theft, or unrelated remote access behavior is present. The code performs only the advertised DoS action.

ClassificationExploit
Model confidence95%
AuthenticationNot required
LanguagesPython
Target softwareJUNG Smart Visu Server
Attack typesDenial of Service
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively sends a crafted request to trigger a server reboot or shutdown, which is the intended exploitation of the vulnerability, not merely detection.

exploits/multiple/webapps/52536.py:29-33exploits/multiple/webapps/52536.py:60-63exploits/multiple/webapps/52536.py:88

Requirements

  • Target running JUNG Smart Visu Server <= 1.1.1050exploits/multiple/webapps/52536.py:10
  • Network access to the target serverexploits/multiple/webapps/52536.py:111

Observed behavior

  • Sends an unauthenticated payload withheld request to payload withheld or payload withheldexploits/multiple/webapps/52536.py:30-33exploits/multiple/webapps/52536.py:66-73
  • Interprets HTTP 200, 302, 301, timeout, or connection error as likely successful exploitationexploits/multiple/webapps/52536.py:78-98
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Url
Payload withheldEndpoint targeted by the PoC to trigger a server reboot.exploits/multiple/webapps/52536.py:31
Url
Payload withheldEndpoint targeted by the PoC to trigger a server shutdown.exploits/multiple/webapps/52536.py:32
Http Method
Payload withheldHTTP method used to send the exploit request.exploits/multiple/webapps/52536.py:66
Review boundaries

What the analysis did not establish

  • Only the single Python file was analyzed; no external libraries or referenced resources were inspected.
  • The script's actual effectiveness or safety was not tested; classification is based solely on code intent.
  • Only the supplied text file was reviewed; no binary files or external resources were inspected.
  • The review does not assess whether the target endpoints are actually vulnerable or whether the script works as intended.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

SumatraPDF 3.5.2 - Remote Code Execution

ExploitDB exploitPublished 2026-04-30
ScannerCVE-2026-259611 file

EDB-52535

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a Flask-based server that impersonates the SumatraPDF update endpoint. It serves a fake update response pointing to a dummy executable, but does not perform the network interception (MITM) required to redirect the victim's traffic. It only demonstrates the server-side component of the attack chain.

Backdoor review

No backdoor observed in reviewed code

The PoC is a straightforward Flask server that serves a fake update response and a dummy executable to demonstrate the CVE-2026-25961 MITM attack. It contains no concealed behavior, persistence mechanisms, credential theft, or unrelated payloads. The dummy payload is explicitly labeled as fake and non-functional.

ClassificationScanner
Model confidence95%
AuthenticationNot required
Languagespython
Target softwareSumatraPDF
Attack typesman-in-the-middle
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The code only provides the server-side component that would serve a malicious update and payload if the attacker already has a MITM position. It does not include any capability to intercept, redirect, or poison the victim's network traffic, which is explicitly stated as a prerequisite. Therefore, it cannot independently exploit the vulnerability; it only demonstrates or validates the server-side behavior.

exploits/multiple/webapps/52535.py:25-26exploits/multiple/webapps/52535.py:36-37

Requirements

  • Attacker must achieve a network position to intercept or redirect the victim's update check request (e.g., DNS spoofing, rogue AP, router compromise).exploits/multiple/webapps/52535.py:25-26
  • Victim must manually trigger the update check (Help → Check for updates) and click Install.exploits/multiple/webapps/52535.py:32-33

Observed behavior

  • Starts a Flask web server on port 5000.exploits/multiple/webapps/52535.py:87
  • Serves a fake update-check-rel.txt response containing a URL to a malicious installer hosted on the same server.exploits/multiple/webapps/52535.py:58-65
  • Serves a dummy executable (malicious_installer.exe) when requested.exploits/multiple/webapps/52535.py:67-76
  • Generates a non-functional dummy payload if the payload file does not exist.exploits/multiple/webapps/52535.py:52-56
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Dummy Payload
Payload withheldThe script generates a non-functional dummy executable if none exists, explicitly instructing the user to replace it with a real payload. This is standard PoC behavior and not a backdoor.exploits/multiple/webapps/52535.py:52-56
Network Service
Payload withheldThe server listens on all interfaces to serve the fake update and payload, as required for the MITM demonstration. No hidden listeners or reverse connections are established.exploits/multiple/webapps/52535.py:78-87
Review boundaries

What the analysis did not establish

  • The artifact does not include the MITM, DNS spoofing, or traffic redirection component required to complete the attack.
  • The dummy payload is non-functional and explicitly marked as a placeholder.
  • Only the supplied Python script was reviewed; no external dependencies or referenced resources were inspected.
  • The dummy payload is not executed or analyzed beyond its textual representation in the script.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

NiceGUI 3.6.1 - Path Traversal

ExploitDB exploitPublished 2026-04-30
ExploitCVE-2026-257321 file

EDB-52534

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

A Python script that sends a multipart POST request with a user-supplied malicious filename (containing path traversal sequences) to a NiceGUI target, attempting to write an arbitrary file outside the intended upload directory.

Backdoor review

No backdoor observed in reviewed code

The supplied PoC is a straightforward Python script that demonstrates a path traversal vulnerability in NiceGUI. It reads a local payload file and sends it as a multipart POST request with a user-supplied malicious filename. No backdoor, deceptive payload, or concealed harmful behavior is present.

ClassificationExploit
Model confidence95%
AuthenticationNot required
LanguagesPython
Target softwareNiceGUI
Attack typesPath TraversalArbitrary File Write
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively sends a crafted request containing a path traversal payload to a target server with the intent of writing a file to an arbitrary location. It is not merely checking for the vulnerability; it performs the exploitation action.

exploits/multiple/webapps/52534.py:19-20exploits/multiple/webapps/52534.py:42-44exploits/multiple/webapps/52534.py:52-57

Requirements

  • Target NiceGUI application must use the unsanitized FileUpload.name property in a file save path (e.g., UPLOAD_DIR / file.name).exploits/multiple/webapps/52534.py:20
  • Attacker must provide a target URL, a malicious filename (e.g., '../app.py'), and a local payload file.exploits/multiple/webapps/52534.py:23-24

Observed behavior

  • Reads a local payload file into memory.exploits/multiple/webapps/52534.py:36-37
  • Constructs a multipart file upload request where the filename field is set to the attacker-controlled malicious filename.exploits/multiple/webapps/52534.py:42-44
  • Sends the crafted POST request to the target URL and prints the HTTP response status and a snippet of the response body.exploits/multiple/webapps/52534.py:52-70
  • Prints post-exploitation guidance suggesting checking the target filesystem or waiting for application reload if a critical file was overwritten.exploits/multiple/webapps/52534.py:75-78
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Url
Payload withheldExample target URL shown in usage docstring; user supplies the actual target.exploits/multiple/webapps/52534.py:23-24
Filename
Payload withheldExample malicious filename demonstrating path traversal; user controls this argument.exploits/multiple/webapps/52534.py:23
Filename
Payload withheldExample local payload file to upload; user controls this argument.exploits/multiple/webapps/52534.py:23
Review boundaries

What the analysis did not establish

  • The script does not verify whether the file was actually written to the target filesystem; it only infers success from the HTTP response code.
  • The script relies on the target application using a specific vulnerable pattern (unsanitized FileUpload.name in a file path), which may not be present in all NiceGUI deployments.
  • Only the supplied text file (52534.py) was reviewed; no external libraries, dependencies, or referenced resources were inspected.
  • The review does not assess whether the target application is actually vulnerable or whether the exploit works as claimed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Cybersecurity AI (CAI) Framework 0.5.10 - Command Injection

ExploitDB exploitPublished 2026-04-30
ExploitUnlinked1 file

EDB-52530

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that directly calls the vulnerable find_file function from the CAI framework with malicious arguments to execute arbitrary OS commands (whoami, touch) as a proof of concept.

Backdoor review

No backdoor observed in reviewed code

The supplied Python script is a straightforward proof-of-concept (PoC) for CVE-2026-25130, a command injection vulnerability in the CAI Framework. It imports a vulnerable function, calls it with crafted arguments to demonstrate OS command execution, and includes a commented-out reverse shell example. No concealed, deceptive, or operator-directed harmful behavior beyond the stated exploit demonstration is present.

ClassificationExploit
Model confidence95%
AuthenticationNot required
LanguagesPython
Target softwareCybersecurity AI (CAI) Framework
Attack typesOS Command Injection
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script directly invokes the vulnerable function with crafted arguments to execute arbitrary OS commands (whoami, touch), which constitutes active exploitation rather than mere detection or scanning.

exploits/multiple/webapps/52530.py:33-44exploits/multiple/webapps/52530.py:51-63

Requirements

  • The vulnerable cai-framework library (<= 0.5.10) must be installed and importable.exploits/multiple/webapps/52530.py:22
  • The /tmp directory must exist and be readable/writable for the PoC payloads.exploits/multiple/webapps/52530.py:38

Observed behavior

  • Imports and calls find_file with a malicious args parameter containing '-exec whoami ;' to execute the whoami command.exploits/multiple/webapps/52530.py:33-44
  • Calls find_file with a malicious args parameter containing '-exec touch /tmp/CAI_CVE_2026_25130_PROOF ;' to create a marker file.exploits/multiple/webapps/52530.py:51-63
  • Contains a commented-out reverse shell payload using bash and /dev/tcp.exploits/multiple/webapps/52530.py:70-72
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Vulnerability Demonstration
Payload withheldThis is the intended purpose of the PoC as described in the header and comments.exploits/multiple/webapps/52530.py:33-42exploits/multiple/webapps/52530.py:51-61
Commented Out Code
Payload withheldThe script explicitly warns against running it outside an isolated environment, and the code is inactive.exploits/multiple/webapps/52530.py:70-72
Review boundaries

What the analysis did not establish

  • The analysis is based solely on the provided Python script; the actual vulnerable library code (cai.tools.reconnaissance.filesystem) is not included in the evidence.
  • The script's execution and effectiveness are not verified; classification is based on the stated and visible intent of the code.
  • The review is based solely on the supplied text of the PoC script. The imported module 'cai.tools.reconnaissance.filesystem' is not included in the evidence, so its behavior cannot be independently verified. The script's execution is not observed; the analysis is static.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

Craft CMS 5.6.16 - RCE

ExploitDB exploitPublished 2026-04-29
ExploitCVE-2025-324321 file

EDB-52525

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that exploits CVE-2025-32432 in Craft CMS by poisoning a PHP session file and triggering a Yii deserialization gadget chain to achieve unauthenticated remote command execution.

Backdoor review

No backdoor observed in reviewed code

The provided Python script is a proof-of-concept exploit for CVE-2025-32432 targeting Craft CMS. It performs session poisoning and deserialization to achieve remote code execution as described. No backdoor, deceptive payload, or concealed operator-directed harm was observed. The script's behavior is consistent with its stated purpose.

ClassificationExploit
Model confidence95%
AuthenticationNot required
Languagespython
Target softwareCraft CMS
Attack typesremote code executiondeserializationsession poisoning
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively executes a multi-step attack chain (session poisoning, deserialization trigger) to achieve arbitrary command execution on a remote target, which is the definition of an exploit.

exploits/multiple/webapps/52525.py:1exploits/multiple/webapps/52525.py:13-16exploits/multiple/webapps/52525.py:68-88exploits/multiple/webapps/52525.py:91-120

Requirements

  • Target running vulnerable Craft CMS version (<= 3.9.14, <= 4.14.14, <= 5.6.16)exploits/multiple/webapps/52525.py:8
  • Valid Asset ID on the target Craft CMS instanceexploits/multiple/webapps/52525.py:27-65
  • Target must write PHP session files to /tmp/sess_<PHPSESSID>exploits/multiple/webapps/52525.py:94

Observed behavior

  • Obtains a PHP session ID from the targetexploits/multiple/webapps/52525.py:143-151
  • Optionally brute-forces a valid Asset ID by sending deserialization payloads to /actions/assets/generate-transformexploits/multiple/webapps/52525.py:37-65
  • Poisons the PHP session file by sending a GET request with PHP code in a query parameterexploits/multiple/webapps/52525.py:75-76
  • Triggers deserialization via a crafted JSON payload to /actions/assets/generate-transform, causing PhpManager to include the poisoned session file and execute the injected commandexploits/multiple/webapps/52525.py:97-120
  • Prints the server response, which may contain command outputexploits/multiple/webapps/52525.py:180-181
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Command Execution
Payload withheldThe script injects a user-supplied command into a PHP session file and triggers its execution via deserialization, which is the core exploit mechanism.exploits/multiple/webapps/52525.py:75
Network Request
Payload withheldAll network requests are directed solely at the user-specified target URL, with no external or unrelated endpoints contacted.exploits/multiple/webapps/52525.py:38exploits/multiple/webapps/52525.py:76exploits/multiple/webapps/52525.py:97
Session Handling
Payload withheldThe script obtains a session ID from the target to construct the path to the poisoned session file, a necessary step for the exploit.exploits/multiple/webapps/52525.py:144-145
Review boundaries

What the analysis did not establish

  • Only the exploit script is analyzed; external libraries (requests, urllib3) and their behavior are not inspected.
  • The script's effectiveness depends on server-side conditions (session file path, asset ID existence) that are not verified in the evidence.
  • No network traffic or execution output is captured; analysis is based solely on static code review.
  • The review is limited to the provided Python script text. No external dependencies (e.g., the 'requests' library) were analyzed.
  • The script's behavior depends on the target environment; no runtime analysis was performed.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

HAX CMS 24.x - Stored Cross-Site Scripting (XSS)

ExploitDB exploitPublished 2026-04-29
ExploitCVE-2026-227041 file

EDB-52526

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

Python script that authenticates to a HAX CMS instance and uploads an HTML file containing arbitrary JavaScript, demonstrating stored XSS (CVE-2026-22704).

Backdoor review

No backdoor observed in reviewed code

The PoC script is a straightforward stored XSS demonstration against HAX CMS. It logs in, uploads an HTML file containing a user-chosen JavaScript payload, and prints the URL. No backdoor, deceptive payload, or concealed operator-directed harm is present.

ClassificationExploit
Model confidence95%
AuthenticationRequired
Languagespython
Target softwareHAX CMS
Attack typesstored cross-site scripting
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script actively uploads a crafted HTML file containing executable JavaScript to a vulnerable endpoint, which constitutes exploitation of the stored XSS vulnerability rather than mere detection or scanning.

exploits/multiple/webapps/52526.py:1-135

Requirements

  • Low-privileged authenticated user credentials for the target HAX CMS instance.exploits/multiple/webapps/52526.py:14
  • Target HAX CMS version <= 24.x with file upload functionality enabled.exploits/multiple/webapps/52526.py:9

Observed behavior

  • Logs into the target HAX CMS instance using provided credentials.exploits/multiple/webapps/52526.py:69-78
  • Generates an HTML file containing a JavaScript payload (alert, cookie, or custom).exploits/multiple/webapps/52526.py:30-63
  • Uploads the malicious HTML file to the /files/upload endpoint.exploits/multiple/webapps/52526.py:86-92
  • Outputs the URL where the uploaded HTML file can be accessed to trigger the XSS.exploits/multiple/webapps/52526.py:98-101
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Payload Generation
Payload withheldThe payload is user-supplied and intended for XSS demonstration; no hidden or malicious payload is embedded.exploits/multiple/webapps/52526.py:30-63
Authentication
Payload withheldAuthentication is required for the exploit; credentials are passed as command-line arguments, not hardcoded or exfiltrated.exploits/multiple/webapps/52526.py:66-82
File Upload
Payload withheldThis is the core exploit step; no additional files or payloads are uploaded.exploits/multiple/webapps/52526.py:86-101
Review boundaries

What the analysis did not establish

  • Analysis based solely on static code review; script was not executed.
  • No network traffic or runtime output captured to confirm successful exploitation.
  • Relies on assumptions about the target application's upload endpoint and file serving path.
  • Only the supplied text file was reviewed; no binary analysis was performed.
  • The script's behavior depends on the target server's response; no server-side analysis was conducted.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.

ExploitDB

LangChain Core 1.2.4 - SSTI/RCE

ExploitDB exploitPublished 2026-04-29
ExploitCVE-2025-686641 file

EDB-52514

Analysisdeepseek-v4-pro:cloud ·

Technical assessment

The artifact is a Python script that exploits CVE-2025-68664, a serialization injection vulnerability in LangChain Core. It constructs a malicious payload with an 'lc' key, uses the vulnerable dumps() function to serialize it, and then uses load() to deserialize it, which instantiates a PromptTemplate containing a Jinja2 Server-Side Template Injection (SSTI) payload. The script then invokes the template to achieve arbitrary command execution (RCE).

Backdoor review

No backdoor observed in reviewed code

The PoC demonstrates the claimed CVE-2025-68664 SSTI/RCE vulnerability in LangChain Core. It constructs a malicious serialized payload, deserializes it, and triggers command execution via Jinja2 template injection. All actions are consistent with a standard exploit demonstration; no concealed backdoor, unrelated payload, or operator-directed harm is present.

ClassificationExploit
Model confidence100%
AuthenticationNot required
Languagespython
Target softwarelangchain-core
Attack typesdeserialization of untrusted dataserver-side template injectionremote code execution
Evidence & reasoningClassification basis · observed behavior · safety review
Technical evidence

Classification basis and observed behavior

Classification basis

The script's primary purpose is to exercise the CVE-2025-68664 vulnerability by constructing a malicious payload, exploiting the unsafe deserialization, and executing an arbitrary system command ('id'). This constitutes an exploit.

exploits/multiple/webapps/52514.py:1-47

Requirements

  • The target application must use a vulnerable version of langchain-core (< 0.3.81 or >= 1.0.0, < 1.2.5) and call load() on attacker-controlled data.exploits/multiple/webapps/52514.py:9-13

Observed behavior

  • The script constructs a dictionary payload containing an 'lc' key with a 'constructor' type that targets langchain_core.prompts.prompt.PromptTemplate.exploits/multiple/webapps/52514.py:23-34
  • The payload's template is a Jinja2 expression that uses Python's built-in import function to execute the system command 'id'.exploits/multiple/webapps/52514.py:30
  • The script serializes the payload using the vulnerable dumps() function.exploits/multiple/webapps/52514.py:37
  • The script deserializes the payload using load(), which instantiates the malicious PromptTemplate object.exploits/multiple/webapps/52514.py:40
  • The script invokes the deserialized PromptTemplate's format() method, which triggers the SSTI and executes the 'id' command.exploits/multiple/webapps/52514.py:43-44
Safety-review evidence

Behaviors behind the backdoor verdict

Observables

Exploit Mechanism
Payload withheldThe payload uses a constructor type with id ['langchain_core', 'prompts', 'prompt', 'PromptTemplate'] and a Jinja2 template that calls payload withheld. This matches the described CVE-2025-68664 serialization injection.exploits/multiple/webapps/52514.py:23-34
Command Execution
Payload withheldThe injected template executes the 'id' command, a benign system information command commonly used in PoCs to demonstrate RCE.exploits/multiple/webapps/52514.py:30
Environment Variable
Payload withheldA dummy environment variable is set to demonstrate potential secret access, but it is not exfiltrated or used for harm. The comment on line 19 explains its purpose.exploits/multiple/webapps/52514.py:19-20
Review boundaries

What the analysis did not establish

  • The analysis is based solely on the static source code of the exploit script. The script was not executed, and its effectiveness against a live target is not verified.
  • The evidence does not include the target application or environment, so the exact context of the vulnerability (e.g., how user input reaches the load() function) is inferred from the CVE description and script comments.
  • Only the single Python file was reviewed; no external dependencies or network resources were inspected.
  • Binary files were flagged but not analyzed; none were present in this artifact.
Model interpretation

This review is limited to the supplied PoC code and context. It does not assert that the code works or is safe to execute.