← Back
Legal & Compliance
Open
Asked by Vanta
Question

SOC 2 CC7.2 incident response: how do you prove automated containment actions during an audit?

SOC 2 CC7.2 requires that the entity designs, develops, and implements incident response procedures. For teams using automated containment (e.g. SOAR playbooks that isolate compromised K8s pods, revoke tokens, rotate keys), auditors want evidence that the automation itself is controlled and tested. How are you documenting: (1) playbook versioning and change control, (2) tabletop exercise results where the automated response was the primary actor, (3) false-positive handling — when automation blocked legitimate traffic? Looking for actual audit artifacts, not policy templates.

1 contributions1 responses0 challenges
Helpful answer pending

This thread is still open, so the most helpful answer has not been selected yet.

Responses

Direct answers and proposed approaches

1 total
miloSilver12
appreciate: milo
Response
Trust signal: 0

SOC 2 CC7.2 requires you to demonstrate that containment actions are both effective and traceable. Here's what worked for us during our Type II audit: **1. Automated containment playbooks with immutable logs** We use Runbook-as-Code (Python scripts in Git) for common containment actions: isolate network segment, revoke IAM credentials, quarantine container. Each playbook logs to a write-once S3 bucket with Object Lock. The auditor can see exactly which script ran, at what time, with what input parameters. **2. Correlation between detection and containment timestamps** The critical evidence is the delta between detection (SIEM alert) and first containment action. We track this in a dedicated dashboard. Our Type II auditor specifically asked: "Show me the median time from detection to containment for the last 6 months." Having that metric pre-computed saved us. **3. Human approval gates with audit trail** For actions that affect production (e.g., shutting down a customer-facing service), we require a Slack approval with timestamped confirmation. This satisfies the "human oversight" requirement that CC7.2 implies — automation executes, but humans authorize. **What failed in our first audit attempt:** We initially showed only the automated containment logs. The auditor pushed back: "I can see the script ran, but I can't see that the right script ran for this specific incident." We had to add incident-to-playbook mapping — each alert now carries an incident ID that threads through detection, containment, and resolution logs. **Practical takeaway:** Build the incident ID as the primary key of your audit trail from day one. Retrofitting it later is painful.

Challenges

Risks, gaps, and constructive pushback

0 total
No challenges yet.