generated: '2026-08-26' method: searched source: https://safe.security/security/ name: SAFE Security vulnerability disclosure url: https://safe.security/security/ http_status: 200 program: published: true type: responsible-disclosure named_policy: Responsible Disclosure Policy policy_url: null policy_note: >- The Security Statement refers readers to "our Responsible Disclosure Policy" for more details, but the phrase carries NO hyperlink anywhere on the page and no such document is reachable from the public site. The policy is named but not published. contact: email: vulnerability_report@safe.security obfuscated: true obfuscation_note: >- The address is not rendered as text - it is a Cloudflare email-protection link whose visible content is an SVG image. Neither a human copy-pasting nor a crawler reading the DOM text gets an address; it has to be decoded from the /cdn-cgi/l/email-protection hex payload. That is a real reporting-friction finding for a security company. security_txt: false security_txt_note: >- No RFC 9116 /.well-known/security.txt is served on any SAFE host, so an automated scanner following the standard path finds nothing. safe.security returns the WordPress homepage shell for that path (soft 404, HTTP 200) and the safeone.ai application hosts return CloudFront AccessDenied. Publishing a security.txt carrying Contact, Policy and Preferred-Languages would be a small, high-value fix. bug_bounty: false bug_bounty_note: >- No HackerOne, Bugcrowd or Intigriti program was found, and no bounty, reward or safe-harbour language appears on the Security Statement. safe_harbour: false disclosure_terms: coordinated: true embargo: indefinite-until-approved quote: >- "You are prohibited from publicly disclosing the details of the reported vulnerability, until the issue has been fixed and you have received prior written approval from SAFE to do so. In the event you fail to strictly comply with the foregoing confidentiality requirement, we reserve the right to institute appropriate legal proceedings." assessment: >- Strict coordinated disclosure with no stated deadline, no safe harbour, and an explicit legal threat. There is no maximum embargo period, so a reporter has no guaranteed date on which they may publish. This is materially less researcher- friendly than the SOC 2 / ISO 27001 posture on the same page would suggest. security_program_published: url: https://safe.security/security/ pdf: https://safe.security/wp-content/themes/safe/assets/security/security-statement.pdf covers: - Secure coding practices (input validation, output encoding, cryptographic practices, session management, access control, authentication and password management, error handling and logging, communication security, system configuration, database security, file management, memory management) - Continuous SAST and DAST as part of DevSecOps - Manual and automated pre-release business-logic testing by an internal security team - Daily vulnerability assessment scans and continuous log/alert monitoring on the AWS production environment - Periodic patch management and access review - Configuration assessment of all PaaS services - Third-party vendor risk assessment using SAFE's own platform, under a 4-tier Vendor Management Policy - Security awareness and secure-coding training; background verification of new recruits evidence: - url: https://safe.security/security/ status: 200 kind: disclosure page - url: https://safe.security/.well-known/security.txt status: 200 kind: soft-404 - WordPress homepage shell, not a security.txt - url: https://us.safeone.ai/.well-known/security.txt status: 403 kind: CloudFront AccessDenied gaps_for_provider: - Publish /.well-known/security.txt on safe.security and the safeone.ai hosts. - Publish the Responsible Disclosure Policy the Security Statement already cites by name. - De-obfuscate or supplement the reporting address so it is machine-readable. - Consider a stated embargo ceiling and a safe-harbour clause.