Test Your VDP Reporting Channel From Contact to Owner
Run an authorised vulnerability disclosure reporting-channel drill. Check discovery, receipt, human ownership and backup with evidence for every handoff.
Ernest Bursa
To test a vulnerability disclosure reporting channel, send a clearly labelled synthetic report through your authorised public route. Check that an external sender can find it, receive a usable receipt, reach a responsible human and get a reply. Then exercise the agreed backup and retain evidence for each handoff.
A working address is only the beginning. On 6 November 2026, Scott Helme described the difficulty of notifying website operators about compromise. His team reported finding 15,579 infected sites; he says most notifications sent manually go unanswered. Those are his team’s observations, not an industry response-rate study.
You can test a narrower question in your own organisation: does your published reporting route lead to someone who accepts the next action? The exercise below gives you a record you can inspect, repair and repeat.
Decide what this drill must prove
This drill checks the path from public contact to human ownership. It uses harmless content and requires no vulnerability, exploit or simulated emergency.
Agree these details before sending anything:
- The public route and assets you are authorised to exercise.
- An approved external sender and a known internal coordinator.
- The primary owner, backup and ordinary queue they should use.
- The test window, expected response window and stop conditions.
- Where you will keep the evidence and who will retest failures.
Notify relevant service providers before exercising their route. Publishing a contact address does not itself authorise security testing, as RFC 9116 explains.
Set the timing to match your published policy and actual staffing commitments. If you have no stated response promise, agree a window for this exercise and label it accordingly. There is no universal 24-hour deadline in this drill.
Also decide how the recipient would hand a genuine compromise report to your incident owner. Helme’s account concerns compromised websites; a vulnerability disclosure programme (VDP) often receives reports of possible weaknesses. RFC 9116 distinguishes vulnerability response from incident response. You need a route between them without treating every vulnerability report as an incident. Your incident response plan should identify the receiving owner.
Start where an outside reporter starts
Ask a colleague to begin at your public website, signed out of internal tools. Use a normal external browser and an approved external mailbox, so the test cannot quietly rely on employee access or internal mail routing.
Have them find the contact route without giving them a private link first. Record the starting URL, route found, elapsed time and any login barrier. Open the policy and submission page as an outsider would.
If you publish /.well-known/security.txt, check its contact and expiry information. RFC 9116 requires Contact and Expires fields and warns about incorrect or inaccessible contact resources. The file helps someone discover a route; it does not provide the whole response process.
If the route is missing or blocked, record that failure before helping the tester continue. A workaround can reveal later problems, but it cannot turn failed discovery into a pass. If you still need the underlying policy and intake process, start with the VDP setup guide.
Send a clearly labelled synthetic report
The message should identify the exercise immediately and ask for next steps you can verify. The coordinator knows the plan; the recipient can see that no real vulnerability or compromise is alleged.
Adapt this template to your organisation and agreed test date:
Subject: [Authorised channel drill] Synthetic report DRILL-2026-11-08-A
This is an authorised reporting-channel exercise for our organisation.
No vulnerability or compromise is alleged.
Please confirm receipt with the report reference, identify the team
or person responsible for the next step, and reply through the normal
reporting route.
Exercise coordinator: [known internal contact]
Synthetic identifier: DRILL-2026-11-08-A
No sensitive data is included.
Use the published form or address. Retain the submission time, synthetic identifier and any visible result. Do not attach a payload, use stolen credentials, plant malware or invent a live emergency. If a form cannot accept an explicitly synthetic report, agree a safe way to exercise it with its owner first.
Keep the synthetic identifier separate from any reference your intake system creates. Recording both helps you follow the same exercise through submission, receipt and replies.
Check receipt, ownership and the reply
Delivery, receipt and human ownership are separate checkpoints. Passing one does not prove the next.
SMTP acceptance is one transport signal; it does not prove the report reached the queue or the reporter got a usable receipt. A usable receipt gives the reporter confirmation they can retain and refer to, such as an acknowledgement email or a report reference on a receipt page. An HTTP success response alone does not establish that a report was saved or routed.
Human ownership needs separate evidence: a person accepts responsibility for the next action in the ordinary queue. A team can be the responsible owner, but someone in that team must actually accept the work. An automatic acknowledgement or assignment label does not show that anyone has read it.
Ask the owner to reply through the normal route. Confirm that the response arrives at the external mailbox, refers to the same report and allows the sender to continue the conversation. Reply once from that mailbox and check that the continuation reaches the owner.
This follows the approach in the UK NCSC vulnerability disclosure toolkit: respond promptly, involve the responsible product or service person, keep the finder informed and communicate the outcome. The drill is our proposed way to check those handoffs, not an NCSC-mandated test.
Exercise the backup without disrupting real reports
Agree that the primary owner will step aside from the exercise while coverage for real reports and incidents stays intact. Then ask the backup to find and accept the same reference through your documented fallback.
Do not disable production email, forwarding or security controls to make the simulation realistic. If you want to test after-hours handling, coordinate staffing explicitly. The useful question is whether the backup can act without relying on the absent person’s inbox or memory.
If there is no independently usable fallback, record that finding. Do not improvise by sending the report to an unrelated third party. Write down who will establish the fallback and when you will exercise it again.
Keep an evidence record and retest failures
Copy this worksheet into your internal exercise record. For each checkpoint, add the observed result, evidence location, timestamp and any fix owner or retest date.
| Checkpoint | Evidence to retain | Failure to fix |
|---|---|---|
| Public discovery | Starting URL, contact or policy URL, time, access barriers | An outsider cannot find or open the advertised route |
| Submission | Synthetic identifier, channel, sent time, visible result | Bounce, form error or blocked submission |
| Receipt | Usable external acknowledgement or reference and timestamp | A technical success signal leaves the sender without confirmation |
| Human ownership | Responsible person or team, acceptance and next action | Automation responds while nobody accepts the work |
| Two-way reply | External reply received and continuation linked to the reference | Replies disappear or enter the wrong thread |
| Backup | Agreed absence simulation, fallback used, backup acceptance | Only the primary owner can reach or understand the report |
| Closure | Failed steps, fix owners, retest dates, exercise-closed reply | Fixes are assumed to work without another test |
Record separate elapsed times for discovery, receipt and human acceptance. Mark any missing checkpoint as incomplete rather than averaging it into an overall score. Send an explicit exercise-closed reply so nobody mistakes the remaining thread for unfinished security work.
Retest failed steps after fixing them. Repeat after contact or identity changes, domain moves, form updates or routing changes. Any regular cadence you choose is a local operating decision, not an RFC requirement.
Keep synthetic data in the record and redact addresses or internal routes before sharing it publicly. A successful run shows what happened on the tested path under those conditions. It does not establish remediation quality, round-the-clock coverage, delivery across every mail provider or readiness for bulk reports.
Keep the report and owner together in Kit
Kit Security can keep a report reference, acknowledgement and assignment in the same workflow. Automatic assignment depends on your configuration and an available assignee; an acknowledgement still needs an external delivery check and human follow-through.
Use the worksheet to test your actual intake configuration from the public starting point. Finish with a usable receipt, an accepted next action and a backup who can continue the work. Where the chain breaks, give that specific failure an owner and a retest date.
Related articles
Try Kit for 30 days.
Hiring, security reports, and training in one account, for teams where none of it is a full-time job. Free for 30 days, card required. Cancel before it ends and you pay nothing.
Get started free