Supported languages:
Bounty rules:
We accept vulnerability reports only if the vulnerability was previously unknown to the VK team.
The types of vulnerabilities eligible for bounties are listed in the "Rewards" section at the end of the rules for the Bug Bounty program rules.
The bounty amounts shown in the description are for reference only.
The applicability and amount of a bounty may depend on the severity of the problem, novelty, likelihood of use, environment, and/or other factors.
The VK security team makes a bounty decision for each message individually.
Any vulnerabilities not listed in the "Rewards" section are paid for at the discretion of the program owner.
Bug Bounty program's scope:
Domains:
This program covers services and products related to or managed by VK but hosted outside the VK infrastructure: fresh and non-integrated acquisitions, domains not mentioned in other programs, various cloud services and hosted off-infrastructure solutions.
Bugs identified at YCLIENTS, YPLACES, INTICKETS, Taxmon are accepted as informational and are not paid for.
Bugs identified at adblogger.vk.com, *.taxmon.ru, welcome.vk.company are accepted as informational and are not paid for.
Bugs detected on demo stands that do not demonstrate a threat to the production project infrastructure (for example access to product source code via Nexus, Tarantool EE, or similar) are accepted as informational and are not paid for.
When testing any functionality that affects other users on
https://expert.vk.com/ you must use these user IDs - 147414, 312717, 327685.
Out of scope domains and projects:
Projects: Umschool, Uchi.RU, AliExpress Russia.
*.my.games and all services related to My.GAMES or MGVC and their partners.
*.vkcs.cloud, *.algoritmika.org, *.yaklass.ru, *.umschool.net, *.code-class.ru, *.summerstage.ru , *.vk-stadium.ru, *.bargipsy.com, *.tetrika-school.ru
Important:
The program awards only critical server-side vulnerabilities if the vulnerability compromises infrastructure (e.g. RCE, SQLi, LFR, SSRF, etc.) or data outside the scope of the VK controlled project (e.g. personal information) via a server-side vector.
Client-side vulnerabilities (XSS, CSRF) and bugs related to business logic, including privilege escalation within the product, are accepted without reward.
MitM and local attacks, open redirects, insufficient session validation, handling cookies after logout, etc. are not accepted unless additional vectors are defined (e.g., the ability to steal a session token via a remote vector for open redirects).
Bugs identified on demo stands, dev infrastructure, domains used for training, delegated, externally hosted domains and partner services are accepted as informational and are not paid for.
0-day/1-day vulnerabilities may be considered as a duplicate within several weeks after vulnerability details publication.
Bug reports submitted by current or former employees (up to one year from the end of employment) of the VK Group are accepted without payment.
When testing RCE, SQLi, LFI, LFR, SSTI it is allowed to use only MINIMALLY possible POC for proof (sleep, accessing /etc/passwd, curl).
Publishing or disclosing bug report and/or any details associated with VK products without approval from VK's information security team is prohibited. We reserve the right to refuse any request for public disclosure of the report.
Vulnerability testing should only be done on your own accounts.
MitM and local attacks, open redirects, insufficient session validation, handling cookies after logout, etc. are not accepted unless additional vectors are defined (e.g., the ability to steal a session token via a remote vector for open redirects).
Limitations on the Bug Bounty program's scope:
It is recommended to limit all scanning tools to 10 requests per second.
We do not accept or review:
- Bug reports from vulnerability scanners and other automated tools;
- Disclosure of information that is not confidential, for example, the version of a product;
- Disclosure of information about a user that is public, for example, a user's nickname;
- Bug reports based on the version of a product/protocol (e.g. TLS version) without demonstrating the actual presence of a vulnerability;
- Bug reports about a missing security mechanism/current best practice (e.g. missing - CSRF token, framing/clickjacking protection) without demonstrating an actual impact on the security of users or the system;
- Messages about published and unpublished SPF and DMARC policies;
- Cross-site request forgery leading to logout (logout CSRF);
- Vulnerabilities in partner products or services, unless Mail.Ru or VK.com users/accounts are directly affected;
- Security of rooted, jailbroken, or otherwise modified devices and applications;
- Ability to reverse engineer an application, or the lack of binary protection;
- Open redirection vulnerabilities are accepted only if a security impact is identified, such as the possibility of stealing an authorization token;
- Injecting unformatted text, audio, images, or video into a server response outside of the user interface (for example, into JSON data or an error message), unless doing so replaces the user interface, changes the behavior of the user interface, or results in other negative consequences;
- Same site scripting, reflected downloads, and similar attacks with questionable impact;
- CSP-related bug reports;
- IDN homograph attacks;
- XSPA (scanning the IP addresses/ports of external networks);
- Excel CSV formula injection;
- Scripting in PDF documents;
- Attacks that require full access to a local account, browser profile or physical access to the device;
- Attacks based on scenarios where a vulnerability in a third-party site or application is required as a prerequisite and is not demonstrated;
- Theoretical attacks without proof of feasibility;
- Denial of service (DoS) vulnerabilities, for example - sending a large volume of requests or data (flooding);
- Ability to send a large number of messages;
- Ability to send spam or a malware file (for example, registration or password recovery spam);
- Disclosure of information through external links not controlled by Mail.Ru or VK.com (for example, Google dorking of private protected areas of robots.txt);
- Disclosure of unused or properly restricted JS API keys (for example, an API key for an external map service);
- Ability to perform an action not available through the user interface and without identified security risks;
- Vulnerabilities associated with the use of phishing and other social engineering techniques;
- Disclosure of /metrics, /status, htaccess or similar without a demonstrated information security threat (for example, disclosure of private API methods, tokens);
- Blind SSRF vulnerabilities without demonstrating a threat to the service's information security in the report;
- EXIF metadata in images;
- SSRF vulnerabilities that involve sending requests via rentgen*.smailru.net, snipster.*.go.mail.ru, mpr*.m.smailru.net, rs-proxy*.i.smailru.net or other proxies specifically designed to protect against SSRF;
- Vulnerabilities that disclose only user accounts but not passwords or other personal data.
We consider bug reports as informational if:
- The vulnerability discloses information about hacked accounts of external users for Mail.Ru or VK.com services;
- The vulnerability is identified in a service independently hosted by the user (Mail.Ru\VK Cloud hosting network, hosting of gaming team resources, hosting of student or laboratory work for educational projects, etc.).
The program awards only critical server-side vulnerabilities if the vulnerability compromises infrastructure (e.g. RCE, SQLi, LFR, SSRF, etc.) or data outside the scope of the VK controlled project (e.g. personal information) via a server-side vector.
Rewards:
| Vulnerabilities | Maximum Bounty |
|---|
| Remote code execution (RCE) | 250 000 ₽ |
| Server-side Injections (SQLi or an alternative) | 125 000 ₽ |
| Access to and work with local files (LFR, RFI, XXE) without jail / chroot / file type restrictions | 125 000 ₽ |
| RCE/LFI in the dev infrastructure / isolated or virtualized process | 50 000 ₽ |
| Non-blind SSRF (with the ability to read the response text), except for dedicated proxies* | 50 000 ₽ |
| Blind SSRF, except for dedicated proxies* | 20 000 ₽ |
| Server-side vulnerability involving disclosure (e.g. memory leaks / IDORs) of critical or highly sensitive application data | 10 000 ‑ 125 000 ₽ |
| Server-side vulnerability involving disclosure (e.g. memory leaks / IDORs) of protected personal data or sensitive client information | 10 000 ‑ 100 000 ₽ |
| Server-side vulnerability involving disclosure (e.g. memory leaks / IDORs) of sensitive application or infrastructure data / organizational role privilege escalation | 10 000 ‑ 100 000 ₽ |
| Admin/support authentication bypass | 100 000 ₽ |
| Blind XSS in the admin/support interface | 75 000 ₽ |
*SSRFs are paid for only when demonstrate access to the project's Internal Network.
Self-XSS, XSS specific to non-common browsers (e.g. IE), blocked CSPs and other vectors without proven script execution are generally accepted without reward. Subdomain takeovers are considered under the same severity/conditions as cross-site request forgery (CSRF).
Detailed error output, local installation path, phpinfo() output, performance counters, etc. are not considered confidential; such messages are usually accepted without payment of a bounty. Messages about disclosure of software versions are not accepted.