Sharing
Sharing & Public Pages
A completed scan does not have to stay in your history. Hand someone a read-only link, list it in a public directory, publish a stable per-host report, drop a live badge in a README, or diff two runs to show what changed. Every surface reads the same stored scan, and every one has an explicit privacy boundary drawn in code.
- Share expiry, in days
- 7/30/90
- Hosts in the public browse
- 500
- Scans per host trend
- 30
Overview
VulnRadar stores each scan once. The public-facing surfaces are different views over that one row, each gated by its own flag on the scan_history record. Two flags do most of the work, and they are deliberately independent of each other:
share_tokenturns a scan into a read-only link at/shared/<token>. A live token also gets ashare_publicly_listedflag that decides whether the link shows up in the browsable Public Scans directory.is_publicdecides whether the scan feeds the per-host cache behind/host/<hostname>and the "all public hosts" browse. A scan can be listed in the directory and never appear on its host page, or the reverse. Neither flag reads the other.
Everything below is one of those views. Creating and revoking share links, listing them, and building badges are all session-only actions, the same as webhooks: a logged-in user, never a Bearer API key. The read-only viewer pages, the directory, and the host report need no account at all.
The Public Scans directory
The Public Scans directory is an unauthenticated, paginated, most-recent-first list of scans someone chose to make discoverable. Each entry links straight to its read-only report. A row appears only when all three conditions hold on the underlying scan.
| Condition | What it means |
|---|---|
| share_token IS NOT NULL | The scan has an active share link. Revoke the link and it drops out of the directory too. |
| share_publicly_listed = true | The per-share listing flag is on. This is separate from is_public and the host page. |
| not expired | share_expires_at is null or still in the future, the same live check the viewer uses. |
What sets the listing flag
When a genuinely new share link is created, its listing flag is resolved once. An explicit choice on the request wins; otherwise it falls back to the scan owner's account default (share_publicly_listed_by_default, whose own column default is true). The fallback is scoped to the scan owner, not whoever clicked Share, because it is the owner's identity and scans that end up listed. If that lookup fails, the share fails closed to not listed: a database hiccup should never be the reason a scan lands in a public directory its owner never opted into. So in practice, sharing a scan lists it by default unless you or your account settings say otherwise.
After creation, the listing status only changes through the explicit per-share toggle in the Shared reports row menu. That toggle requires an active share link (a scan with no link returns 400) and is scoped to the same owner or team-write check as sharing itself.
What a viewer can see
Sharing a scan already exposes its findings, so a read-only viewer of your own link sees the whole report: findings, the severity summary, response headers, tags, any note you attached, and who ran it (name, avatar, staff role, and profile badges). That is the same detail you see on your own dashboard, minus the private per-finding remediation state, which stays with the owner.
The viewer page also carries a subdomain-discovery panel in read-only mode. It only ever shows an already-cached snapshot: an anonymous visitor has no session or key to kick off a new discovery job, and it is not their scan to spend rate-limit budget on.
Per-host reports
/host/<hostname> is a public, crawlable report for a host, in the spirit of securityheaders.com or SSL Labs. It reads the host_reputation cache and shows a host's most recent public scan: the danger score, severity counts, findings, response headers, SSL grade, DNS, ports, software inventory, auto-tags, and an AI summary if the owner generated one.
The danger-score trend
The report also charts the host's risk score over time. The reputation cache only keeps the single latest scan, so the trend reads scan_history directly under the same privacy scope: is_public = true and a completed status. It matches the host with a boundary-aware pattern, so example.com and www.example.com both count while evil.example.com does not, and plots up to the 30 most recent scans oldest to newest.
Assets and attack surface
Two logged-in views sit next to History and Public Scans and roll per-scan data up to the host and portfolio level.
Assets: hosts you have scanned
The Assets page groups your scans by host, most-recently-scanned first, with each host's latest safety verdict and scan count. A scope toggle switches what you are looking at:
- My scans (the default) reads your own
scan_history, grouped by host, and honors your plan's retention window the same way your history list does. GitHub repo scans are left out, since they have their own view. - All public hosts (
?scope=all) reads only thehost_reputationcache, which is public-scan-only by construction. It exposes nothing that is not already public on a host page, capped at 500 hosts so the list stays a browse and not a dump. Your own private scans never appear here.
Attack surface: verified domains
The Attack surface page is your verified-domain portfolio, and it is where you add and verify a domain. Verifying a domain proves you own it: a verified apex covers every subdomain beneath it and unlocks active probing, authenticated scans, and subdomain discovery across those assets. The verification backend already existed inside the profile Developer tab; this surfaces it as a first-class, portfolio-level view of the assets you own and monitor.
Security badges
A badge is an SVG image that reads "Secured by VulnRadar" alongside the site’s latest grade (A+ through F) and scan date, linking through to the full read-only report. The safe/caution/unsafe verdict is still there, as the colour of the right-hand half: green, yellow, red. The grade leads because every comparable free tool grades a site on that scale, and "caution" on a README is something an owner removes rather than keeps. The key property: a badge is tied to a URL, not to one frozen scan. Every time you scan that URL again, the same embedded image updates on its own. Paste the snippet once and leave it.
Generating and embedding
Pick a scan on the Badge page and it issues an auto-updating token for that URL. The token is idempotent per URL, so asking twice returns the same one and your embed code never has to change. Copy one of the three snippets:
<a href="https://vulnradar.dev/shared/YOUR_TOKEN" target="_blank" rel="noopener noreferrer" style="display: inline-block;"><img src="https://vulnradar.dev/api/v3/badge/YOUR_TOKEN" alt="Secured by VulnRadar" style="border: 0;"/></a>[](https://vulnradar.dev/shared/YOUR_TOKEN)The image lives at /api/v3/badge/<token> and the click-through goes to /shared/<token>. The image is cached for an hour, so a fresh scan may take up to that long to show through a CDN.
User scope vs global scope
A badge's scope decides which scan it reflects, and you can toggle it any time without regenerating the token.
| Scope | Resolves to |
|---|---|
| user | The default. Only your own scans of that URL. The badge updates when you rescan it, and never leaves your account. |
| global | The newest scan of that URL by anyone, but only if that scan's owner marked it public. Lets the badge stay current when other people scan the URL too. |
Revoking a badge stops its token from resolving without touching the underlying scan history, and an expired or revoked token renders a neutral "Link Expired" image instead of a verdict.
Comparing two scans
The Compare page diffs two scans of the same host so you can see what actually moved between runs: which findings appeared, which you closed, and which have been sitting there the whole time.
How the diff works
Pass two scan ids as a and b. Both are required, and both must belong to you: a scan id on another account resolves to 404, so there is no way to diff scans you do not own. The two are ordered oldest first, and their findings are matched by title into three buckets:
| Bucket | What it means |
|---|---|
| added | Present in the newer scan but not the older one: findings that appeared. |
| removed | Present in the older scan but not the newer one: findings that were fixed or are gone. |
| unchanged | Present in both: findings that have persisted across the two runs. |
The UI only offers hosts with two or more scans on record, since a diff needs a before and an after. With exactly two, it runs the diff immediately; with more, you pick the pair and the older one becomes the base. The selection lives in the URL (?a=&b=), so a comparison is a link you can share with a teammate.