PT Sandbox
Bounty Range
Up to $6,000
external program
PT Sandbox
Company: Positive Technologies
The PT Sandbox Bug Bounty Program aims to identify and confirm vulnerabilities that may lead to bypassing malware detection mechanisms, executing malicious code outside of an isolated environment, compromising sandbox infrastructure, and reducing protection effectiveness against targeted attacks.
PT Sandbox is a locally deployable sandbox for analyzing unknown and complex malware, including zero-day exploits, ransomware, and targeted attacks. The product uses a combination of static and behavioral analysis, correlation rules, network sensors, ML algorithms, and OS monitoring mechanisms. Vulnerabilities in PT Sandbox can lead to missed malicious objects, isolation environment breaches, and compromise of system components and related infrastructure.
At the time of program launch, access to product test environments is provided on a limited basis.
Extended access will be provided later, as infrastructure and support procedures become ready.
Within this program, vulnerabilities are classified by bypass levels, reflecting the depth and criticality of impact on PT Sandbox analysis and isolation mechanisms.
| Vulnerability Name | Description | Attack Vector Requirements |
|---|---|---|
| Sandbox Escape | A file submitted for behavioral analysis changes execution context and begins executing code in the host OS or hypervisor level. The bypass result should be the ability to execute arbitrary code on the host system or disrupt its operation. | The researcher has the ability to submit files for analysis from outside by any supported method. Manual behavioral analysis of files in the product interface is permitted. |
| Vulnerability Name | Description | Attack Vector Requirements |
|---|---|---|
| Malicious Document Bypass | Undetected execution of malicious code in user context using documents of popular office applications (MS Office, LibreOffice, PDF, etc.). Conditional execution of malicious behavior outside PT Sandbox is permitted. | The file must run on user click, not require additional software, and be intended for a supported platform and OS. |
| Hiding Active Element in Office Document | A malicious element correctly implements dangerous behavior but is not detected by PT Sandbox static analysis. Possible techniques include OLE objects, macros, DDE, ActiveX, JavaScript, OpenAction, external data sources, and Office add-ins. | The file must be a valid document, run on user click, and be intended for a supported platform and OS. |
| Hiding Popular Format from Behavioral Analysis | A valid executable file or document is incorrectly classified as inactive and is not passed to a virtual environment for analysis (e.g., EXE or office document). | The file must be intended for a PT Sandbox supported platform and OS. |
| Vulnerability Name | Description | Attack Vector Requirements |
|---|---|---|
| Malicious Executable Bypass | Undetected execution of malicious code using executable files (EXE, ELF) in user context. Conditional execution of malicious behavior outside PT Sandbox environment is permitted. | The file must run on user click, not require additional conditions, and be intended for a supported platform and OS. |
| Web Interface and API: Authorization Bypass | Vulnerabilities in the interface and public API, including authorization bypass, XSS based on submitted content for analysis, insecure deserialization, and Path Traversal during file uploads. | The researcher can submit files through the web interface, Public API or email integrations, and has network access to available system services. |
Note: Vulnerabilities that do not lead to real risk (for example, theoretical or without exploitation confirmation) may be rejected or evaluated as "informational" without monetary compensation.
Reward amounts are described in the table below:
| Severity Level | Reward Amount |
|---|---|
| Critical | ₽300,000 – 500,000 |
| High | ₽150,000 – 300,000 |
| Medium | ₽50,000 – 150,000 |
| Low | ₽0 – 50,000 |
Compensation can only be paid for attack scenarios that can be reproduced on installations of officially supported product versions with all available updates. Reports of vulnerabilities in unsupported versions are also accepted, but compensation for such vulnerabilities is not guaranteed.
The vulnerability severity level is determined during triage and report confirmation, taking into account impact on product security.
The final decision on the vulnerability severity level is made by the product security team.
All interested researchers aged 18 and over may participate in the program.
Researchers aged 14 to 18 may participate in the program only with written consent from parents or a legal guardian.
Current employees of Positive Technologies and former employees who left less than 3 years ago may participate in the program but cannot claim compensation.
Comply with the rules established by Positive Technologies in its vulnerability disclosure program and the rules of The Standoff 365 Bug Bounty platform.
Comply with confidentiality rules. It is prohibited to access another user's data without consent, modify and destroy it, or disclose any confidential information accidentally obtained during vulnerability research or demonstration. Intentional access to this information is prohibited and may be considered illegal.
Maintain communication with the security team, submit reports of identified vulnerabilities formatted according to requirements, and provide feedback if specialists have questions about the report.
Not disclose vulnerability information. The right to publish information about a found vulnerability remains with Positive Technologies.
Vulnerability disclosure is only permitted if there is a fix and a publicly registered CVE/BDU identifier.
A bug bounty hunter may request vulnerability report disclosure – PT is obligated to initiate the process of coordinating the registration of a vulnerability identifier.
Positive Technologies does not pay compensation for:
Reports from security scanners and other automated tools;
Disclosure of non-secret information (software names or versions, technical parameters and system metrics, etc.);
Information about IP addresses, DNS records, and open ports;
Issues and vulnerabilities based on the version of the product used, without demonstration of their exploitation;
Vulnerabilities whose exploitation is blocked by security tools, without demonstration of security tool bypass;
Reports of insecure SSL and TLS ciphers without demonstration of their exploitation;
Reports of absent SSL and other best current practices;
Vulnerabilities whose information was previously submitted by other competition participants (duplicate reports);
Zero-day or 1-day vulnerabilities whose information was obtained by the security team from public sources;
Brute force attack vulnerabilities, if the report does not describe a method with significantly higher efficiency than direct brute force.
Launched December 26, 2025