Patch Management Policy Template for SOC 2
A SOC 2 patch management policy sets the cadence and remediation SLAs for applying security updates across cloud infrastructure, containers, application dependencies, and employee workstations. Auditors test this policy under AICPA CC7.1 and CC8.1 by inspecting vulnerability scan logs, Dependabot/Snyk alerts, and remediation timestamps against your stated SLAs.
Free Markdown Template
Copy or download the patch management policy
Replace bracketed [PLACEHOLDERS] with your actual stack
# Patch and Vulnerability Management Policy
**Document Owner:** [Engineering / Infrastructure Lead]
**Approved By:** [CTO / Security Lead]
**Effective Date:** [YYYY-MM-DD]
**Review Cadence:** Annual
**Applies To:** Production cloud infrastructure, containers, source-code dependencies, and employee endpoints
---
## 1. Purpose
This policy establishes a repeatable process and clear service-level agreements (SLAs) for identifying, prioritizing, testing, and deploying security patches across [Company Name]'s infrastructure, applications, and endpoints.
## 2. Scope and Asset Coverage
Patch management applies to all in-scope components supporting [Company Name]'s production service:
- Cloud hosts, managed databases, and Kubernetes/container base images ([AWS / GCP / Azure])
- Application code libraries and third-party packages ([GitHub Dependabot / Snyk / Trivy])
- Production SaaS and CI/CD runners
- Company-managed employee workstations ([macOS / Windows via MDM])
## 3. Continuous Vulnerability Detection
1. **Code & Dependencies:** Automated dependency and container image scanning runs on every pull request and at least daily on default branches via [GitHub Dependabot / Snyk / AWS Inspector].
2. **Infrastructure & Cloud:** Automated cloud posture and host vulnerability scanning runs continuously via [AWS Inspector / GCP Security Command Center / Wiz].
3. **Endpoints:** [MDM Provider] monitors workstation OS versions daily and enforces automatic security updates.
## 4. Severity Classification and Remediation SLAs
Vulnerabilities are prioritized using CVSS v3.1/v4.0 severity adjusted for production reachability and exploitability:
| Severity | CVSS Range | Remediation SLA | Verification |
|---|---|---|---|
| **Critical** | 9.0 – 10.0 (or actively exploited KEV) | [24 – 72] hours / [7] days | Re-scan confirms finding closed |
| **High** | 7.0 – 8.9 | [14 / 30] days | Re-scan confirms finding closed |
| **Medium** | 4.0 – 6.9 | [60 / 90] days | Next scheduled release / scan |
| **Low** | 0.1 – 3.9 | Best effort / [180] days | Tracked in backlog |
*Note: Adjust SLA numbers to match what your team can consistently meet across a 12-month Type II window.*
## 5. Testing, Deployment, and Emergency Patches
1. **Standard Patches:** Infrastructure and dependency updates follow the standard Change Management Policy (PR review, automated CI tests, staging verification where applicable).
2. **Emergency Patches:** Critical actively exploited vulnerabilities may use an expedited change path with written approval from [CTO / Engineering Lead] and post-deployment verification logged within [24] hours.
3. **Endpoints:** Workstation OS security updates must be installed within [14] days of vendor release; MDM enforces restarts after the grace period.
## 6. Risk Acceptance and Exceptions
If a patch cannot be applied within SLA (for example, no upstream fix exists or the vulnerability is unreachable in production), the owner must document a risk exception in [Jira / Linear / Risk Register] containing:
- Vulnerability ID (CVE) and affected asset
- Compensating controls or reachability justification
- Written approval from [CTO / Security Lead] and an expiration date (maximum [90] days)
## 7. Review
This policy is reviewed annually and updated whenever scanning tools, cloud environments, or SLA commitments change.
Evidence Mapping
What auditors test beyond this policy
A signed policy document only satisfies design intent. During SOC 2 fieldwork, your CPA auditor samples operating evidence against each clause below:
| Policy Requirement | Control | Evidence Your Auditor Requests |
|---|---|---|
| Automated dependency & cloud scanning enabled | CC7.1 | Screenshots or API scans of GitHub Dependabot, container scanner, and AWS Inspector/GCP SCC showing continuous scanning across production repos and clusters. |
| Critical and High vulnerabilities fixed within SLA | CC7.1 | Closed vulnerability tickets or Dependabot PRs from the observation period comparing detected date vs. patched date against Section 4 SLAs. |
| Documented exceptions for overdue CVEs | CC7.1 / CC3.2 | Written risk acceptance records with compensating controls and approval dates for any open High/Critical alert older than your SLA. |
| Workstation OS patching enforced | CC7.1 / CC6.6 | MDM compliance report showing employee laptops running patched OS versions. |
- Setting a '24-hour Critical / 7-day High' SLA in the policy because it sounds strict, then missing it on a noisy npm transitive dependency and earning a Type II audit exception.
- Leaving dozens of dismissed or stale Dependabot alerts unreviewed in GitHub without documenting why they are unreachable or not applicable.
- Testing only application packages while forgetting base Docker images, managed Postgres minor versions, or employee laptops.
Related guide: GitHub SOC 2 report and repository controls auditors test
Frequently Asked Questions
Patch Management Policy FAQ
- What patch management SLAs are realistic for a startup SOC 2?
- A common, auditor-accepted startup baseline is 7 days for Critical (or 48–72 hours for actively exploited CVEs), 30 days for High, and 90 days for Medium. Never write tighter SLAs than your team actually meets, because Type II auditors test your own stated deadlines.
- How do I handle false positives or unreachable CVEs in Dependabot?
- Document a brief triage note (or formal risk exception for Critical/High findings) explaining why the vulnerable code path is not reachable in production, with sign-off from your engineering or security lead and a review date.
- Does SOC 2 require an external vulnerability scanner if we use GitHub Dependabot?
- Dependabot covers code dependencies, not running cloud infrastructure or endpoints. Pair dependency scanning with cloud/container scanning (such as AWS Inspector, Trivy, or Snyk) plus MDM OS monitoring so all layers are covered.
Skip manual fill-in-the-blank edits
Draft this policy directly from your real infrastructure.
Screenata scans your AWS, GCP, GitHub, and identity provider read-only, writes auditor-ready policies tied to real controls, and collects signed evidence automatically.