Security · Restricted bug bounty

High-impact vulnerabilities only.
Demonstrate a real security failure.

We review reproducible RCE, authentication bypass, and other specifically listed high-impact security failures. Reports requiring prior victim compromise, victim interaction, or social engineering are excluded. Read every requirement before testing or submitting.

01 · Overview

Demonstrable impact. Strict scope.

This is a restricted program for reproducible, high- or critical-impact vulnerabilities. Hardening suggestions, speculative attack scenarios, and low-impact findings are not eligible.

We accept reports only when they demonstrate one of the security outcomes listed below on a current, in-scope eVPS deployment. The report must show a real security boundary being crossed, with a complete, reproducible attack path.

In-scope assets: websites, customer control panels, first-party APIs, and provider-managed services operated by Redcluster LTD for eVPS. Customer-managed VPS operating systems, customer applications, hosted domains, and third-party-operated services are excluded. An eVPS IP address, DNS record, or brand reference alone does not establish scope. If ownership is unclear, obtain written confirmation before testing.

Required attacker model: a remote attacker, without prior compromise of the victim's device, mailbox, authenticator, account, or authenticated session, and without requiring victim interaction. Phishing, stolen authentication material, and persuading a victim to perform an action are not eligible prerequisites.

Using your own ordinary customer account is permitted when it demonstrates an independent server-side flaw that grants access to another account, protected data, or privileges you do not hold. Actions already allowed within your own account are not vulnerabilities. The limited MFA-bypass case is defined below.

Only reports meeting every requirement are eligible for reward consideration. Any reward, its type (financial compensation or eVPS.net merchandise), and its amount are determined by Redcluster LTD after validation. Submission, a CVSS score, or a severity label does not create an entitlement to a reward.

Safe testing and disclosure
  • Report promptly. Use only accounts you created and resources you control. Demonstrate cross-account issues between two of your own test accounts; never target another customer.
  • Use the minimum non-destructive proof needed to establish impact. Do not install persistence, deploy malware, pivot to other systems, modify shared infrastructure, or extract databases or secrets.
  • If you unexpectedly encounter third-party data, stop immediately, retain only the minimum redacted evidence needed to explain the issue, and report it. Do not continue accessing or downloading data.
  • Do not use phishing, impersonation, social engineering, physical access, or any testing that requires another person to act on your instructions.
  • Do not brute-force credentials or codes, flood endpoints, send unsolicited messages, or disrupt availability. Ask for written approval before any validation that cannot be performed safely within these limits.
  • Keep the report confidential and allow Redcluster LTD a reasonable opportunity to investigate and remediate before public or third-party disclosure. Personal information provided will not be shared without consent unless disclosure is required by law.
  • Following these expectations and the other conditions in this policy, Redcluster LTD does not intend to pursue legal action in relation to the report itself.
02 · Accepted impact

Only these outcomes qualify for review.

All scope, attacker-model, evidence, and safe-testing requirements apply. A vulnerability name alone is not enough; the resulting impact must be reproducible and assessed by us as high or critical.

Remote code execution
  • Execution of attacker-controlled code or operating-system commands on an in-scope provider-managed system through a remotely reachable vulnerability.
  • Running code inside your own VPS, using an intentionally available console, or exploiting a system you already administer does not qualify.
Authentication bypass or account takeover
  • A server-side flaw that grants an authenticated session or control of a test victim account without the authentication normally required, including a reproducible password-reset or recovery bypass.
  • For an MFA-bypass report only, the test may start with the correct password for your own MFA-enabled test account. It must bypass the enforced second factor from a fresh unauthenticated session, without a valid or stolen OTP, recovery code, trusted-device token, or existing authenticated session.
Major authorization or isolation failure
  • A normal customer account gains administrative privileges or unauthorized control over a different account, VPS, console, backup, or other protected resource through a server-side authorization or isolation flaw.
  • Demonstrate this between your own test accounts or resources. A changed identifier, hidden button, successful HTTP status, or client-side response edit without actual unauthorized access is insufficient.
Serious confidential-data exposure
  • A remote flaw exposes private cross-account data or provider secrets with a demonstrated high or critical security impact, such as private backups, authentication secrets, or sensitive customer records.
  • Public information, usernames, software versions, internal paths, IP addresses, and other non-sensitive metadata do not qualify. Do not retrieve other customers' records to prove scale; use your own test data and stop at the minimum safe evidence.
03 · Out-of-scope findings

Excluded reports will be closed.

Reports that do not meet this policy are not eligible for rewards and may be closed or ignored without acknowledgment, individual explanation, or further correspondence. Repeated submissions do not change eligibility.

Prior compromise or local access

  • Scenarios requiring physical or remote access to the victim's computer, phone, browser profile, local storage, files, operating system, or network interception position.
  • Malware, malicious extensions, an already compromised mailbox or account, and copied or stolen passwords, cookies, sessions, API keys, TOTP seeds, OTPs, backup codes, or trusted-device tokens. The narrowly defined MFA-bypass test above is the only password-known exception.
  • Taking authentication material from your own account and presenting it as stolen from someone else, without a separate in-scope flaw that exposes it.

Victim interaction and social engineering

  • Phishing, impersonation, support manipulation, fake login pages, or persuading a person to disclose credentials or codes.
  • Attacks requiring a victim to click an attacker-supplied link, visit a page, open or upload a file, paste code, run a command, install software, approve a prompt, change settings, disable MFA, or perform another attacker-induced action.
  • Physical attacks against staff, property, infrastructure, or data-center facilities.

TOTP, sessions, and account hardening

  • Using, replaying, or sharing a valid or stolen TOTP code; possession of a TOTP seed; accepted clock windows; and OTP reuse observations without the independent server-side MFA bypass defined above.
  • Actions performed after already obtaining full access to the victim account: changing email or password, disabling MFA, creating tokens, or accessing resources the session is authorized to use.
  • Missing reauthentication prompts, session duration or invalidation preferences, password-complexity rules, email-verification preferences, MFA-enrollment observations, and missing login or account-change notifications, without an eligible independent exploit.

Headers, TLS, email, and configuration

  • Missing or imperfect security headers, CSP, HSTS, cookie flags, noreferrer/noopener, CORS observations, or TLS/certificate/cipher hardening recommendations without an eligible exploit.
  • SPF, DKIM, DMARC, email spoofing, or email-delivery configuration observations.
  • Version banners, public IPs, internal paths, non-sensitive directory listings, OPTIONS or other HTTP methods being enabled, and error messages or stack traces without sensitive exposure.

Low-impact web and browser findings

  • Self-XSS, browser-console demonstrations, client-side-only bypasses, clickjacking, CSRF, open redirects, homograph attacks, tabnabbing, and content or visual spoofing that do not meet the accepted-impact and no-victim-interaction requirements.
  • XSS, cache poisoning or deception, SSRF callbacks, CORS findings, or injection claims supported only by an alert, callback, error, timing change, or theoretical chain, without evidence of an accepted high- or critical-impact outcome.

Scanners and hypothetical claims

  • Unvalidated scanner output, AI-generated or template reports without service-specific proof, generic checklists, and vulnerability claims based only on a product name, version string, public CVE, or automated severity score.
  • Best-practice recommendations, outdated-software observations, speculative chains, or demonstrations limited to a local environment that do not establish exploitability on the in-scope deployment.
  • Requests for payment to reveal details, sales pitches, and reports that omit the affected endpoint, necessary preconditions, reproducible steps, or demonstrated impact.

Availability, guessing, and abuse

  • DoS/DDoS, resource exhaustion, stress testing, brute force, credential stuffing, password or OTP guessing, and missing rate limits or weak CAPTCHA without an independent eligible exploit.
  • User or email enumeration, spam, referral or coupon abuse, pricing complaints, billing disputes, and other commercial-abuse scenarios that do not produce an accepted security outcome.
  • Do not perform disruptive or abusive testing to establish impact, including when another part of the report would otherwise qualify.

Customer systems and external services

  • Customer-managed VPSs, customer websites or applications, browser or operating-system vulnerabilities on victim devices, and services operated by third parties.
  • Generic third-party software or WordPress reports and payment-processor parameter tampering outside the direct eVPS application surface. A dependency issue qualifies only if its actual deployment on an in-scope eVPS system demonstrably produces an accepted outcome under this policy.

Complete exploit chains only

  • Low-impact observations do not qualify on their own. If an observation forms part of a complete exploit chain, submit the chain as one finding and prove an accepted outcome.
  • This does not override the exclusions on assets, prior compromise, victim interaction, or prohibited testing methods.
04 · Before reporting

Evidence first. No speculative submissions.

Provide a complete, reproducible, non-destructive proof that establishes the claimed security boundary failure. Real victim access, destructive exploitation, and bulk data extraction are neither required nor permitted.

Eligibility
  1. You are the first reporter to provide a complete, reproducible report of the issue. Duplicate symptoms, endpoints, or exploit variants with the same underlying cause are treated as one issue.
  2. The affected asset, attacker prerequisites, and demonstrated outcome meet every requirement of this policy. Redcluster LTD validates actual impact and determines eligibility and severity.
  3. You follow the safe-testing and disclosure requirements and do not publicly disclose before Redcluster LTD has had a reasonable opportunity to investigate and respond.
  4. You provide enough evidence to reproduce the flaw safely. A screenshot, scanner score, or statement that an issue could lead to RCE or account takeover is not sufficient by itself.
  5. You test only with your own accounts and permitted resources. No existing victim session, victim cooperation, or third-party compromise may be assumed.
  6. You are not in, or subject to sanctions under, the U.S. SDN list or the EU Consolidated Financial Sanctions list.
Required report contents
  • The exact affected hostname, endpoint, service, and accepted-impact category.
  • The attacker's starting permissions and everything required: accounts, known credentials, session state, MFA state, tokens, configuration, and any dependency on another action or vulnerability.
  • Numbered reproduction steps and a working, minimal proof of concept or request/response sequence. Include relevant timing and environment details.
  • Expected versus actual behavior, the security boundary crossed, and evidence of the resulting unauthorized capability or serious exposure. Distinguish observed results from assumptions.
  • For account or authorization issues, identify your attacker and test-victim accounts and show the action occurring without the test victim's authenticated session or action. For MFA bypass, start with a fresh unauthenticated session.
  • Redacted requests, responses, or logs sufficient for validation, plus contact details. Do not send live passwords, OTP seeds, reusable session tokens, or unnecessary personal information.
05 · Reporting channel

Send your report to our security team.

Submit complete, in-scope reports to the address below, using the subject “[eVPS Security] Impact category – affected endpoint”. Keep each issue and its follow-ups in one thread. Customer support is not an alternative route for excluded reports.

Impact first

Meets every requirement?
Send the complete proof.

Only validated, in-scope reports meeting every requirement are considered for a discretionary reward. Excluded, incomplete, speculative, and duplicate reports are not eligible and may receive no response.

Reporting details

Strict scope, reproducible impact, discretionary rewards.

Reporting channelEmail the security team
Response modelEligible reports only
Reward modelDiscretionary cash or merch
Legal entityRedcluster LTD