Tenable Security Center (WIP)
Reads Tenable Security Center (on-premises, formerly Tenable.sc) and does two things:
- Endpoints. Every host Security Center has scanned is created or enriched as an endpoint, like any other endpoint integration.
- Vulnerabilities. Every finding on those hosts, open and fixed, is kept with its history in the tenant's vulnerability index, apart from the endpoints. Dashboards, risk and remediation reporting read it.
| Type name | tenable_sc |
| Creates | Endpoints |
| Runs on | A sensor with HTTPS reach to Security Center |
| Presence flag | tsc_exists |
| Field prefix | tsc_ |
| Vulnerability index | vulns-<tenant id> |
How it reads Security Center
Everything comes from Security Center's analysis API, which only queries: nothing is changed in Security Center. Each run reads, for the configured repositories:
| What | Security Center's view | Becomes |
|---|---|---|
| Scanned hosts | The IP summary of the cumulative database | Endpoints |
| Open findings | Vulnerability details, cumulative database | Findings in the vulnerability index |
| Fixed findings | Vulnerability details, mitigated database | Findings in the vulnerability index |
Each run reads them whole. Security Center answers a few thousand findings in a few seconds, and reading everything is what lets a run tell which findings are no longer reported at all.
A configured repository that the user cannot read, or that does not exist, fails the run with a message listing the repositories the user can read. If the findings cannot be read, the hosts are still delivered and the findings already kept are left as they are.
Endpoints
Matching
A host is matched to the endpoint it already is by, strongest first:
- Its Security Center host key: the repository and Security Center's host UUID, or the repository and the IP address for a host without a UUID.
- The Nessus agent that scanned it: the same agent the Nessus agents integration records.
- Its name and its IP address together.
- Its name, short or fully qualified.
- Its IP address, only on an endpoint whose name does not say it is another host.
A host known by nothing but an address (no name and no MAC), which is typical of an uncredentialed sweep, enriches the endpoint it matches but never creates one.
What it writes
Security Center is a scanner's view of the host. The shared fields names, fqdn,
domains, os, type and private_ip are written only when empty, so what an
agent on the host reported stays. The MACs the scan saw are added to mac beside
what other sources reported; placeholder MACs that many machines share are never
written. os_vendor is written from Security Center's operating system on every
run. The endpoint's last_seen is its most recent scan, credentialed or not, and
only moves forward.
| Field | Holds |
|---|---|
tsc_host_key, tsc_host_uuid | How Security Center tells the host apart |
tsc_agent_uuid | The Nessus agent behind an agent scan |
tsc_ip, tsc_dns_name, tsc_netbios_name, tsc_mac | The host's addresses and names |
tsc_os, tsc_bios_guid | Its operating system, from Security Center's CPE, and its BIOS GUID |
tsc_repository, tsc_policy, tsc_plugin_set | Where its data lives and how it was last scanned |
tsc_last_auth_scan, tsc_last_unauth_scan | Its last credentialed and uncredentialed scans |
tsc_critical, tsc_high, tsc_medium, tsc_low | Its open findings by severity |
tsc_score, tsc_acr, tsc_aes | Security Center's vulnerability score, asset criticality rating and asset exposure score |
The two last-scan dates are decommissioning evidence: a scan only finds a host that is there.
Vulnerabilities
One document per finding
A finding is one plugin on one host, port and protocol. It keeps the same document for its whole life, open, fixed or back again, so a run never duplicates one.
| Field | Holds |
|---|---|
asset_id | The endpoint the host is. Findings are written after the endpoints are synced, so a host created by the same run is linked at once. A link made by an earlier run stays when this run cannot make one |
source, source_finding_id | tenable_sc, and Security Center's own ID for the finding |
check_id, title, category | What was found: the plugin ID, plugin name and plugin family |
severity, severity_name, source_severity | How bad: 0 Info to 4 Critical |
cve, cvss2, cvss3, cvss3_vector, cvss4, epss | Scores and the CVEs it covers |
exploit_available, exploit_frameworks | Whether it can be exploited |
port, protocol | Where on the host |
summary, solution, description, evidence, see_also | The plugin's synopsis, solution, description, plugin output and references, each cut to between 1,000 and 4,000 characters |
vuln_published, patch_published | Publication dates |
risk_accepted | Whether the finding's risk was accepted in Security Center |
ip, hostname, fqdn, mac, os, host_key, host_uuid | The host as Security Center reported it |
scope_id, scope_name, scope_uuid | The Security Center repository the finding is in |
vendor | What only Security Center reports: vpr, risk_factor, exploit_ease, check_type, xref, plugin_published, plugin_modified, recast_risk, dns_name, netbios_name |
collected_at, updated_at | When the run wrote the document |
The field names are scanner-neutral so that findings from other scanners can share the
index; what is specific to Security Center sits under vendor.
Its lifecycle
| Field | Holds |
|---|---|
state | open, fixed, reopened (fixed once and found again) or removed |
first_seen, last_seen | First and last time a scan found it |
fixed_at | The scan that found it fixed, the most recent time |
reopened_at, reopen_count | When it last came back, and how often |
removed_at | When Security Center stopped reporting it at all |
history | Every change of state with its time, the latest 50 |
A finding is removed when a run that read everything finds it neither open nor fixed.
Its host was removed from Security Center, or its data aged out, without a scan finding
the finding fixed. It is not counted as remediated. Only open and reopened findings in
the repositories the run read, and at a severity the run looked for, can become
removed: switching Include informational findings off does not remove the Info
findings already kept.
Security Center keeps fixed findings for a short while only, often a few weeks. The vulnerability index keeps every finding it has seen, so history builds up from the first run onwards. What was fixed before the first run and has already gone from Security Center cannot be recovered.
Informational findings
Findings of severity Info are mostly inventory and configuration facts, and are usually the large majority of what a scan reports. They are left out unless Include informational findings is on.
Requirements
Network access
| From | To | Protocol | Port |
|---|---|---|---|
| The sensor running the integration | Security Center | HTTPS over TCP | 443 |
Access in Security Center
| Need | Why |
|---|---|
| API keys of a Security Center user | Generated in Security Center under the user's API keys. API keys must be enabled for the organisation |
| That user able to list the repositories | Each run, and the connection test, starts by listing the repositories the user can read, to resolve the configured names |
| That user able to read the vulnerability data of the repositories | Every read is an analysis query scoped to those repositories |
Give the integration a user of its own whose role can only read the repositories it needs. An administrator's keys work, but carry far more than the integration needs.
Confirm the least-privileged Security Center role that can run vulnerability analysis on a repository.
Configuration
| Field | Required | Notes |
|---|---|---|
| Name | Yes | Label for this instance |
| URL | Yes | Security Center, e.g. https://securitycenter.contoso.com |
| Access Key | Yes | The API access key |
| Secret Key | Yes | The API secret key |
| Repositories | No | Comma-separated repository names or IDs that belong to this tenant, matched without regard to case. Empty reads every repository the user can see |
| Cookies | No | For a Security Center behind an access gateway: name=value; name2=value2 |
| Proxy settings | No | http://host:port |
| Ignore SSL | No | Skips certificate validation. Install Security Center's CA on the sensor instead where possible: the API keys are sent on every request |
| Sensor | Yes | Sensor that will run the integration |
Secret Key and Cookies are secrets: once saved, the console shows each as a locked box with Replace (and Remove for Cookies), never the value, and sends it only to the sensor.
When one Security Center serves several organisations, list each tenant's repositories on its own instance. An instance with no repositories listed reads everything its user can see.
Its scan template adds, under the advanced settings:
| Option | Default | Effect |
|---|---|---|
| Vulnerabilities | On | Keeps every finding in the vulnerability index. Off reads the hosts only |
| Include informational findings | Off | Also keeps severity Info findings |
See also
- Nessus agents: the agents behind Security Center's agent scans
- Integrations overview: the shared configuration model