> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tumban.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Coverage

> What Tumban actually analyzed during a scan.

Every triage report includes a `coverage` object that reports which
analysis steps ran successfully. Use it to interpret partial results
honestly.

## Fields

<ResponseField name="profile_scraped" type="boolean" required>
  `true` if Tumban successfully loaded the profile page and extracted
  bio, images, and links.
</ResponseField>

<ResponseField name="external_search_completed" type="boolean" required>
  `true` if the web search for external mentions of the profile
  completed.
</ResponseField>

<ResponseField name="social_links_checked" type="integer" required>
  Number of links Tumban traversed during analysis. Counts every URL
  followed from the profile bio or surfaced via referrer cross-reference
  — not only social profiles, despite the name. A value of `0` means the
  profile had no traversable outbound links.
</ResponseField>

<ResponseField name="blocked_by_login" type="string[]">
  URLs Tumban could not analyze because they required login. Treated as
  neutral; not used to inflate the risk score.
</ResponseField>

<ResponseField name="referrer_domains_received" type="integer">
  Number of referrer domains you supplied with the scan via
  `metadata.referrer_domains` (an array of domain names). Tumban does
  **not** read analytics platforms on your behalf — this field counts
  only what you passed in. When you do not supply `referrer_domains`,
  this is `0`.
</ResponseField>

<ResponseField name="referrer_domains_after_filter" type="integer">
  Referrer domains remaining after Tumban filters out search engines and
  analytics platforms.
</ResponseField>

<ResponseField name="referrer_matches_found" type="integer">
  Number of the referrer domains you supplied that Tumban corroborated
  against the profile's external web-search results — i.e. a supplied
  referrer domain also appeared among the URLs found mentioning the
  profile elsewhere on the web. This is **not** a match against a
  prohibited-platform list.
</ResponseField>

## When you'll see partial coverage

A scan that experienced one or more step failures (slow page, login
wall, transient model error) today comes back with `status: "completed"`
and the partial-coverage details recorded inside the `coverage` object
itself — `social_links_checked` may be lower than expected, individual
URLs may show up under `blocked_by_login`, and so on. Inspect `coverage`
before drawing conclusions about edges; the `recommendation` and
`risk_score` are still meaningful, but they were produced from a
narrower input set.

Scans that experienced step failures still come back as
`status: "completed"` — the `coverage` object is the canonical record
of what ran and what was skipped (see [Status values](/reference/status)).
Read from `coverage` to detect partial pipelines; do not gate
partial-handling logic on the `status` field.