Skip to main content

Workflow

Triage & Remediation

A scan tells you what is wrong. This is what happens next: you triage each finding, and when you need a human, you talk to us. Remediation tracking gives every finding a status, an assignee, and a due date that carry over to the next rescan of the same target. Support tickets open a threaded conversation with staff from the Contact page, on any plan including free.

Remediation statuses
5
Findings per bulk update
200
Plan can open a ticket
Free

Overview

VulnRadar splits the after-scan workflow into two things a scan report cannot do for you. The first is deciding what to do about each finding and remembering that decision: remediation tracking. The second is asking a human when a scan alone is not enough: support tickets. Both are private to you: remediation is per-user and never appears on a shared or public view of a scan, and a ticket is visible only to you, staff, and any teammate you explicitly loop in.

Remediation tracking is separate from the accuracy feedback on a finding (confirmed, false positive, not applicable). That feedback tunes the global detection confidence model; remediation records what you have done about the finding and nothing else. The two controls sit next to each other on an open finding but never touch the same data.

The HTTP endpoints below live under https://vulnradar.dev/api/v3/, the same base as the rest of the app. The support ticket UI lives on the Contact page under the Support Ticket option.

Remediation tracking

Open any finding on your own scan and you get a Your remediation tracking control: a status, and once the status leaves Open, a note, an assignee, and a due date. It is stored per finding, per user, and it is remembered the next time you scan the same target.

Statuses

Five states, in the order they appear in the control. Open is the default and is never stored as a row; the other four persist.

Remediation statuses, from lib/scanner/remediation.ts
Stored valueControl labelIn the list
openOpenNo badge. Open is the implicit default, so no row is stored: setting a finding back to Open deletes any row it had.
in_progressIn progressBrand-coloured badge. Counts as active work, so the row keeps full opacity in the list.
fixedFixedSuccess-coloured badge. The row dims slightly so still-open findings stand out.
accepted_riskAccepted riskMuted badge (shown as Accepted). Dims in the list like Fixed.
wont_fixWon't fixMuted badge. Dims in the list like Fixed.

Assignee, note, and due date

Once a finding is anything other than Open, three optional fields appear:

  • Note (max 2000 chars): free text, for a release number or an internal ticket reference.
  • Assignee (max 120 chars): a free-text field with a datalist that suggests people you share a team with. Solo users just type a name; the field never forces a real account.
  • Due date: a target or SLA date. When it is in the past and the finding is still open work (not Fixed, Accepted risk, or Won't fix), the finding shows an Overdue badge. A due date on a finding you have already closed out never nags.

Surviving rescans

This is the part that makes tracking worth doing. A remediation row is keyed on the finding, not on the scan that surfaced it. The finding id is deterministic: the same check firing against the same URL always produces the same id, which is the same identity the compare, diff, and regression-alert features already rely on.

text
finding_id = <checkId>--<fnvHash(scannedUrl)>

// Deterministic: the same check against the same URL always
// yields the same id, so a remediation row is keyed on the
// finding, not on the scan_history row that surfaced it.
remediation key = (user_id, finding_id, finding_url)

So a finding you marked fixed on Monday still reads fixed when you rescan the same host on Friday, even though that is a brand new scan row. The status is attached only on the owner's own result-load paths (scan status and history); the public /shared/[token] and /host/[hostname] views never read it, so your private tracking never leaks onto a link you share. If the underlying table has not been migrated yet, the read fails soft to an empty map and the result page still loads.

Bulk actions

Setting one finding at a time is fine for a couple of them, but a fresh scan can surface dozens. The findings list on your own scan has an opt-in Select mode: the list stays clean (no checkboxes) until you turn it on, and only then do row checkboxes and a sticky bulk bar appear. Pick a status, optionally an assignee and a due date, and apply it to everything you selected, up to 200 findings per request.

JSON
// POST /api/v3/scan/remediation/bulk
{
  "items": [
    { "findingId": "hsts-missing--9f2c1a", "findingUrl": "https://example.com" },
    { "findingId": "csp-missing--9f2c1a",  "findingUrl": "https://example.com" }
  ],
  "status": "in_progress",
  "assignee": "alex"
}

Remediation API

One route handles the single-finding lifecycle, plus a bulk variant. All of it is session-gated and owner-scoped: there is no way to read or write another user's remediation state.

Remediation endpoints under /api/v3
MethodEndpointWhat it does
POST/scan/remediationSet or update one finding's status, note, assignee, and due date. status "open" clears the row.
GET/scan/remediationList your remediation rows, optionally filtered by ?url= and/or ?findingId=. Up to 500, newest first.
DELETE/scan/remediationClear one finding's remediation (back to Open). Requires ?url= and ?findingId=.
POST/scan/remediation/bulkApply one status, plus an optional assignee and due date, to up to 200 findings at once.
POST
/scan/remediation

Set a finding's remediation status

Record what you have done about one finding: its status, plus an optional note, assignee, and due date. Upserts on (user_id, finding_id, finding_url), so calling it again for the same finding updates the row rather than duplicating it. This is the owner's private tracking, separate from the accuracy feedback that feeds the global confidence model.

Request Body

JSON
{
  "findingId": "hsts-missing--9f2c1a",
  "findingUrl": "https://example.com",
  "status": "in_progress",
  "note": "patched in release 4.2, ticket VR-118",
  "assignee": "alex",
  "dueAt": "2026-09-01"
}

Response
200 OK

JSON
{
  "ok": true,
  "remediation": {
    "status": "in_progress",
    "note": "patched in release 4.2, ticket VR-118",
    "assignee": "alex",
    "due_at": "2026-09-01",
    "updated_at": "2026-08-24T12:00:00.000Z"
  }
}
Session-gated on the signed-in user's cookie. There is no Bearer-key equivalent: this is a UI-facing owner action, not part of the public API surface.
Owner-scoped by construction. Every row is keyed on (user_id, finding_id, finding_url) and every query filters on your own user id, so you can only ever read or write your own remediation state.
status "open" is special: it deletes the row instead of storing one, because Open is the absence of a record. DELETE /scan/remediation?url=...&findingId=... does the same thing.
note (max 2000), assignee (max 120), and dueAt are all optional. dueAt takes a date-input value like 2026-09-01 or a full ISO datetime; an unparseable value is stored as null rather than erroring.

Error Responses

400
Invalid JSON, or a body that fails validation
401
No session cookie
503
finding_remediation table not migrated yet (run the migration)

Support tickets

When a finding needs a person rather than a fix step, open a ticket. It is a tracked, two-way thread with our team, not a fire-and-forget email form. It lives on the Contact page under the Support Ticket option, and it is available on every plan, including free.

Opening a ticket

Sign in, pick the Support Ticket option on the Contact page, and fill in a subject, a category, and a first message. The thread opens immediately and staff reply into the same thread. There is one guard rail: you can hold at most 20 open tickets at a time, so you cannot stockpile them, and a captcha protects creation when Turnstile is configured.

Categories and statuses

Four categories, chosen when you open the ticket: Billing, Scanning, Account, Other. The status then tracks whose turn it is:

Ticket statuses, from lib/support/ticket-constants.ts
Stored valueLabelWhat it means
openOpenBrand new, no staff reply yet.
awaiting_staffAwaiting staffYou (or a shared teammate) replied. The ball is with staff.
awaiting_userAwaiting your replyStaff replied. The ball is with you.
resolvedResolvedMarked done by you or staff. Replying reopens it as awaiting staff.
closedClosedNo further replies accepted. A reply returns 409; open a new ticket to continue.

The flow is automatic: a new ticket is Open, your reply flips it to awaiting staff, a staff reply flips it to awaiting your reply. You can resolve or close your own ticket; staff can move it to any state, including reopening. In a thread, you see your own messages as "You" and staff replies as "Support": a staffer's real name is never exposed to a non-staff viewer, and your identity is exposed only to staff.

Sharing a ticket with a teammate

A ticket can be shared, but narrowly on purpose. Only the owner can share it, only with a specific teammate (someone who shares a team with you), and one person at a time. There is no share-with-the-whole-team button and no way to share with an arbitrary account: the picker only lists people you already share a team with.

  • A shared teammate can read the thread and reply, but cannot resolve, close, reopen, or manage sharing. Those stay with the owner and staff.
  • The teammate sees the ticket in their own list, flagged Shared with you and labelled with who opened it.
  • The owner can revoke a share at any time, which removes that teammate's access to the thread.

Replies and notifications

A staff reply reaches you two ways at once: an in-app bell notification and an email. Both are transactional (you opened the ticket), so they are sent directly and are not subject to marketing notification preferences, and both link straight back to the thread via /contact?ticket={id}. In the other direction, a new ticket or a user reply emails the staff support inbox.

Support ticket API

The same routes the Contact page UI calls. Every one is session-gated; access to a specific ticket resolves to owner, staff, or a shared teammate.

Support ticket endpoints under /api/v3
MethodEndpointWhat it does
GET/support-ticketsList your own tickets plus any shared with you, newest first.
POST/support-ticketsOpen a new ticket (subject, category, message). Any plan.
GET/support-tickets/{id}Fetch a ticket and its full message thread.
POST/support-tickets/{id}Add a reply to the thread.
PATCH/support-tickets/{id}Change status: resolve, close, or (staff) reopen.
GET/support-tickets/{id}/sharesList current shares and the teammates eligible to add.
POST/support-tickets/{id}/sharesShare the ticket with a teammate ({ userId }). Owner only.
DELETE/support-tickets/{id}/sharesStop sharing with a teammate (?userId=N). Owner only.
POST
/support-tickets

Open a support ticket

Create a tracked ticket and its first message. Available on every plan, including free. The only gate is a signed-in session (plus Cloudflare Turnstile when it is configured). The response is the new ticket row; the UI then opens its thread.

Request Body

JSON
{
  "subject": "Scan times out on staging",
  "category": "scanning",
  "message": "Every scan of https://staging.example.com stalls at 90%."
}

Response
200 OK

JSON
{
  "ticket": {
    "id": 128,
    "subject": "Scan times out on staging",
    "category": "scanning",
    "status": "open",
    "created_at": "2026-08-24T12:00:00.000Z",
    "last_message_at": "2026-08-24T12:00:00.000Z"
  }
}
category is one of billing, scanning, account, other. subject is capped at 200 characters and the message at 5000.
Turnstile is verified the same way the public contact form does it, and is a no-op when Turnstile is not configured for the deployment.
You can hold at most 20 open tickets at once (open, awaiting staff, or awaiting your reply). Resolved and closed tickets do not count toward that.

Error Responses

400
Missing subject or message, bad category, or a failed captcha
401
Not signed in
429
Too many open tickets
Finding Remediation and Support Ticket Tracking | VulnRadar