| name | triage-a-security-alert |
| category | code |
| description | Assess a security signal, preserve evidence, and choose a proportionate next action. Use when monitoring, a vendor, a user, or an automated detector reports suspicious identity, endpoint, network, application, or data behavior. |
triage-a-security-alert
Decide what is known, what could be harmed, and what must happen next. An alert is a claim to investigate, not proof of compromise or proof of safety.
Inputs
- Raw alert, detector logic, timestamps, confidence, and source
- Asset, identity, data, environment, and business criticality
- Relevant telemetry, baselines, recent changes, and response authority
- Escalation, containment, privacy, and evidence-handling procedures
Procedure
- Preserve the original alert and record receipt, owner, source timezone, and any automatic actions.
- Check alert integrity, scope, suppression state, detector health, and known duplicates.
- Identify affected identities, assets, data, privileges, geography, and time window.
- Enrich from authoritative logs while recording query, source, retention gap, and collection time.
- Separate observations, benign explanations, suspicious patterns, and unsupported inference.
- Rate potential impact and urgency using the organization's approved criteria.
- Contain only within authority, choosing reversible steps when facts remain incomplete.
- Escalate to incident response when thresholds or uncertainty justify it; assign next actions and evidence owners.
- Close only with a documented explanation, corroborating evidence, and any detection or control follow-up.
- Review false-positive, missed-signal, and telemetry-gap lessons without weakening coverage blindly.
Guardrails
- Do not contact a suspected actor or alter a device before considering evidence loss and safety.
- Do not upload sensitive telemetry to unapproved analysis tools.
- Absence of one log event is not proof that an action did not occur.
- Keep employee, customer, and sensitive investigation details need-to-know.
Done
- Alert provenance, scope, impact, and confidence are recorded
- Containment and escalation decisions name authority and evidence
- Open hypotheses and telemetry gaps have owners
- Closure is reproducible from the retained case record