Securin Exposure Triage

Exposure Triage

Ad-hoc search, filtering, and aggregation of the user's exposures — instances of vulnerabilities, misconfigurations, or findings discovered in their environment. Translate questions like "show me all open critical exposures breaching SLA" into the correct search*Data / aggregate*Data call with proper scoping and deep links.

An exposure is an instance (e.g., "CVE-2024-3400 on host web-01"). A vulnerability is the global CVE record. This skill is about exposures; use securin-cve-enrichment for vulnerability intel.

When to use

• "Show me all open critical exposures"
• "How many exposures are breaching SLA?"
• "Break down exposures by severity and workspace"
• "List exposures older than 90 days"
• "Exposure volume over time" / "new vs closed exposures"
• "Show me exposures tied to CVE-2024-3400"
• "Which exposures are on exposed-to-internet assets?"

Pre-flight

Step 0 — Account preflight (CC-1)

See _shared/account-preflight.md. Resolve account-id(s) and validate access before any query. If the question implies a workspace subset, resolve workspace-ids too.

Before using this skill, read every file in the references folder, including the shared references/_shared/ docs.

Step 0.5 — Detect composite vs source data model

Exposures are affected by the composite flag too — composite accounts have a separate composite-exposure index with compositeExposure.* field paths and use the exposureQuery tool (source accounts use searchExposureData + aggregateExposureData). See _shared/composite-vs-source.md. Cache the flag for the turn — picking the wrong model returns empty results with no error.

Suggested tools

Primary

• searchExposureData — flat list
• aggregateExposureData — bucketed count (TERMS) and time-bucketed histogram (DATE_HISTOGRAM)
• aggregateVulnerabilityTimelineData / searchVulnerabilityTimelineData — historical vulnerability state changes over time (if available for account; fall back to DATE_HISTOGRAM if 404)

Supporting

• getApiFields with entityType: ["EXPOSURE"] — field discovery
• getGroupByFields — valid aggregation dimensions
• getTopValues — enum values for a field
• getDefaultViewForGroupByField — platform's default column set for a grouping
• validateFilter — FQL construction + validation
• getEffectiveAccessWorkspaces — workspace scoping

Deep links (CC-2)

• createDeepLink (preferred) — call once per list/aggregation, plus once per bucket if you need per-bucket links
• getDeepLink — URL for a known exposure-id

See _shared/deep-links.md.

Workflow

Step 1 — Classify the ask

Step 2 — Discover fields if uncertain

getApiFields(entityType=["EXPOSURE"])
getTopValues(field="exposure.status")
getGroupByFields(entityType="EXPOSURE")

Step 3 — Compose FQL

See _shared/fql-grammar.md. Exposure-specific patterns:

exposure.status = 'Open'
exposure.scores.scoreLevel = 'Critical'
"exposure.scores.score" >= 7.0   # numeric — sort uses the same path: exposure.scores.score:desc
"exposure.firstDiscoveredOn" >= "2026-01-01T00:00:00Z"
exposure.remediationTarget.status = 'Overdue'
exposure.assignments.assignedTo.name = 'team:remediation'
exposure.mappedAttributes.vulnerabilityIds = 'CVE-2024-3400'
exposure.mappedAttributes.type like 'Vulnerability'

Cross-entity filters (exposure → vuln / asset):

# Exposures tied to a specific CVE — correct cross-entity field is vulnerabilities.id
vulnerabilities.id = 'CVE-2024-3400'
# ❌ vulnerabilities.vulnerabilityId  → 400 Invalid field
# ❌ vulnerabilities.cveId           → 400 Invalid field

# Exposures tied to exploited vulnerabilities
vulnerabilities.isCisaKEV = true

# Exposures on exposed-to-internet assets (source-model account)
asset.reachability = 'Exposed'

# Same, composite-model account
compositeAsset.reachability = 'Exposed'

Step 4 — Pick and call the right tool

• For time series: try aggregateVulnerabilityTimelineData / searchVulnerabilityTimelineData first. If those return 404, fall back to aggregateExposureData with DATE_HISTOGRAM.
• For per-bucket deep links, run aggregateExposureData to get the buckets, then call createDeepLink once per bucket with the bucket's filter narrowed by the bucket value.

aggregateExposureData — correct request shape

function is an object with a type discriminator. Type-specific parameters (field, size, etc.) go inside the function object:

{
  "filters": "exposure.status = 'Open'",
  "aggs": [{
    "name": "bySeverity",
    "function": {"type": "TERMS", "field": "exposure.scores.scoreLevel", "size": 10}
  }]
}

Common mistakes that cause 400/500:

• ❌ "function": "TERMS" — function must be an object, not a string
• ❌ "apiPath": "..." — key must be "field", not "apiPath"
• ❌ "field": "exposure.workspaceId" — returns 500; use "asset.workspaces.name" for workspace grouping

aggregateExposureData with DATE_HISTOGRAM — time series shape

Different structure from TERMS: nested aggs use "aggs", and interval must be Title Case. isFixedInterval, extendedBounds, and hardBounds are required:

{
  "aggs": [{
    "name": "openedByMonth",
    "function": {
      "type": "DATE_HISTOGRAM",
      "field": "exposure.firstDiscoveredOn",
      "interval": "Month",
      "isFixedInterval": false,
      "extendedBounds": {"min": "2025-11-01T00:00:00Z", "max": "now"},
      "hardBounds": {"min": "2025-11-01T00:00:00Z"}
    },
    "aggs": [{"name": "count", "function": {"type": "COUNT", "field": "exposure.exposureId"}}]
  }]
}

Valid interval values (Title Case only): "Day", "Week", "Month", "Quarter", "Year". ❌ "month", "MONTH", "1M", "MONTHLY" all return 400.

Step 5 — Deep links (CC-2)

Every result table adds a View in Platform column. Every bucket in an aggregation gets its own URL. Call createDeepLink once per reported scope; getDeepLink(exposureId) for individual drill-downs.

Step 6 — Emit response

**Query:** <plain-English restatement>
**Account:** <account name + id>
**Filter:** `<FQL>`

<results table with View-in-Platform links>

[View full list in Platform](<top-level deep link>)

**Observations:**
- <Pattern noted — e.g., "70% of critical exposures are in one workspace">
- <Suggested next step — e.g., "Drill into remediation: use securin-remediation-guidance for the top 3">

Common recipes

"Open critical exposures breaching SLA"

searchExposureData
filter: exposure.status = 'Open'
        AND exposure.scores.scoreLevel = 'Critical'
        AND exposure.remediationTarget.status = 'Overdue'
sort: "exposure.scores.score:desc"

"Break down exposures by severity and workspace"

Run aggregateExposureData once per dimension (two calls):

// Call 1 — by severity
{
  "filters": "exposure.status = 'Open'",
  "aggs": [{"name": "bySeverity", "function": {"type": "TERMS", "field": "exposure.scores.scoreLevel", "size": 10}}]
}

// Call 2 — by workspace (use asset.workspaces.name — exposure.workspaceId returns 500)
{
  "filters": "exposure.status = 'Open'",
  "aggs": [{"name": "byWorkspace", "function": {"type": "TERMS", "field": "asset.workspaces.name", "size": 25}}]
}

Scope guard (CC-3)

• Global CVE details → hand off to securin-cve-enrichment.
• "Am I affected by threat X" → hand off to securin-threat-correlation.
• "How do I fix this exposure" → hand off to securin-remediation-guidance.
• Pure asset questions → hand off to securin-asset-triage.
• Unknown tool needed → fall back to the platform's built-in Securin__search_tools meta-tool to look up the right MCP tool by description.

Edge cases

• Empty result on cross-entity filter — check asset prefix (asset.* vs compositeAsset.*) matches the account's data model.
• Custom statuses — accounts can customize the exposure state machine; use getTopValues(field="exposure.status") to enumerate.
• Very wide time windows — lean on aggregateExposureData (TERMS or DATE_HISTOGRAM) rather than raw listing.
• Tags / custom attributes — exposure.tags, exposure.customAttributes.* are discoverable via getApiFields.

Visual output (CC-4)

When this skill produces aggregated or multi-row data (counts, trends, distributions, comparisons, single-CVE reports), emit a chart/graph/infographic in the Securin brand — multi-series palette (#9C66FF / #7F30FF / #E96001 / #4D268D / #DD639C / …, assigned in order), semantic CHML severity colors (Critical #A60D08 → Info #C5CBD6), Poppins headings on DM Sans body, light theme, and the Securin logo. Use the 10-stop brand purple ramp for heatmaps/sequential scales; gradients are background decoration only — never on chart bars, lines, or slices. Full color system in _shared/brand.md. Offer customization after delivery; never default to a different brand.

References

• Exposure Fields Quickref
• Shared: Account Preflight
• Shared: Composite vs Source
• Shared: Deep Links
• Shared: FQL Grammar
• Shared: Sorting Rules
• Shared: Brand & Visual Communication