- Section
- Disclosure policy
- Ref
- BM-LG-03
- Updated
Coordinated disclosure.
How we tell vendors about the flaws we find, when we publish, and how to report a flaw in BreachModal without legal risk.
- 120 days
- Standard window for a vendor to ship a fix
- 30 days
- When a patch fails and the flaw is exploited
- 5 business days
- For a vendor to acknowledge before we escalate
- 0 quiet fixes
- No verified flaw is ever suppressed
Form BM-VR · vulnerability report
Report a vulnerability.
Found a flaw in BreachModal? Email it to us, with:
- The affected asset, URL or API endpoint
- Steps to reproduce it
- What an attacker could do with it
- How to reach you, and whether you want credit
- Acknowledged
- Within 48 hours, by a person
- Protection
- Safe harbour, §7
- Machine-readable
- security.txt
§1Scope
This policy works in two directions.
- Outbound: when our research, red team or AI-assisted audits find a vulnerability in someone else’s product, this is how we tell the vendor and when we publish.
- Inbound: when you find a vulnerability in BreachModal, this is how to report it and what protection you have.
It follows ISO/IEC 29147 for disclosure and ISO/IEC 30111 for handling.
§2The disclosure timeline
Disclosure is not an ambush. It is a disciplined sprint to protect users before a flaw is weaponised, and every step is dated.
| When | Stage | What we do | What customers receive |
|---|---|---|---|
| Day 0 | Discovery and first contact | We verify a working proof of concept, classify the weakness (CWE) and score it with CVSS v4.0. We contact the vendor confidentially through its published security channel: an RFC 9116 security.txt, a PSIRT portal, or encrypted email. | Defensive detections and virtual-patch rules, with no exploit details, through encrypted channels. |
| Day 5 | Second attempt | If the vendor has not acknowledged receipt or assigned an engineer within five business days, we try again through alternate leadership, regional technical contacts and other verified channels. | Monitoring raised to look for third-party probing or independent rediscovery. |
| Day 10 | Coordinators engaged | Five business days after the second notice, we ask trusted coordinators to open a dialogue: the CERT Coordination Center, CISA’s vulnerability coordination team or the relevant national CSIRT. | Firewall, WAF and runtime rules reviewed against emerging scanning patterns. |
| Day 15 | Defensive advisory threshold | If every verified channel is exhausted without a response, we may publish a limited advisory fifteen business days after first contact. It covers defensive indicators and workarounds only. | Detection playbooks for defence teams and SOC operators. |
| Day 120 | Coordinated release | Once receipt is confirmed, the vendor has 120 calendar days to build, test and ship a permanent fix. When it ships, or when the window lapses without engagement or a verifiable reason, we publish a coordinated advisory with mitigation guidance. | We verify the vendor’s patch and help customers move from temporary rules to the official fix. |
We never distribute exploit code, proofs of concept or weaponised details before a public fix is available.
§3How we reach a vendor
We first use any verified security contact the vendor publishes, such as /.well-known/security.txt or a bug bounty or disclosure portal.
If none exists, we send an encrypted advisory package to security@, secure@, support@ and info@ at the vendor’s domain, with reproducible findings, our analysis and recommended mitigations.
§4No quiet fixes
A verified vulnerability is never kept quiet because a vendor would prefer it, lacks the resources or calls it “by design”.
If a vendor cannot or will not patch, we offer to publish with them, alongside workarounds that work: architectural mitigations, segmentation and defensive rules, so defenders are not left blind. We also publish a factual, dated summary of our communication with the vendor.
§5Faulty or incomplete patches
A patch that misses the root cause or is trivially bypassed gives defenders false confidence. We apply shorter deadlines, set by severity.
| Window | Applies when | Protocol |
|---|---|---|
| 30 days | Active exploitation or imminent threat · critical, CVSS 9.0 and above. The first patch is faulty, incomplete or bypassed, and exploitation has been seen in the wild or weaponised code is circulating. |
|
| 60 days | High severity, partial protection · CVSS 7.0 to 8.9. The faulty patch still stops opportunistic attacks and no exploitation is seen, but a skilled attacker could get round it. |
|
| 90 days | Other regressions and secondary defects · medium. Secondary bugs, logic flaws and non-critical edge cases found after the patch that need deeper work across platforms. |
|
§6Extensions
Extensions beyond these windows are rare and at our technical discretion. A vendor asking for one must show verifiable evidence of complex rework, such as hardware or firmware recertification, multi-vendor coordination or a substantial operating system rebuild, show continuous progress, and confirm that no exploitation has been seen in the wild.
§7Reporting a vulnerability in BreachModal
We hold our own applications, cloud infrastructure and agent systems to the standard we ask of others. If you believe you have found a flaw in any BreachModal application, service or API, send it to security@breachmodal.com. A person will acknowledge it within 48 hours.
Safe harbour
We will not pursue legal action against researchers who test and report in good faith under this policy. We consider testing within these limits authorised under the Computer Fraud and Abuse Act, and we will not bring claims under the DMCA or other anti-circumvention provisions.
| Authorised | Not authorised |
|---|---|
| Non-destructive proof-of-concept testing, checks of authentication flows, probing the limits of model jailbreaks, and read-only checks of permission isolation. | Denial-of-service or volumetric attacks, destroying user data, social engineering or phishing our staff, and any access beyond what demonstrates the issue. |
§8Official advisories
We publish advisories only on our own site, in the advisory index. An advisory carries BreachModal attribution only if it appears there. We do not release threat claims through third-party channels.