Security policy templates

A complete information security policy set: 22 editable templates with owners, review cadences and the tools you actually use. Free to generate and Apache-2.0 licensed, whether or not you are working towards SOC 2.

One policy, or a set?

Most companies start with a single document called the information security policy, and most auditors, customers and insurers eventually ask for more than that. The workable shape is a short top-level policy that states the scope, the objectives and who is accountable, with supporting policies underneath it that go into detail on access, data, engineering and the rest.

Splitting them is not bureaucracy for its own sake. Each policy ends up with a different owner and a different review rhythm: the engineering lead maintains change management, the person who runs laptops maintains endpoint security. One 40-page document has one owner, which in practice means it has none.

The 22 policies

Every policy below is a real template on this site. Follow any of them to read the full text before you generate your own copy.

Governance

The policy that sets scope and authority, and the rules that apply to everyone.

PolicyWhat it coversTypical owner
Information Security PolicyThe top-level policy: scope, objectives, roles, and the authority the rest of the set hangs fromSecurity Owner
Acceptable Use PolicyWhat staff may and may not do with company systems, accounts and dataSecurity Owner
Risk Assessment and Management PolicyHow risks are identified, scored, assigned an owner and treatedSecurity Owner

Access and identity

Who can reach what, proven by approvals and reviews rather than by memory.

PolicyWhat it coversTypical owner
Access Control PolicyWho gets access to what, how it is approved, and how it is reviewed and removedSecurity Owner
Authentication and Password PolicyPassword rules, MFA, and how credentials and keys are issued and storedSecurity Owner
Asset Management PolicyThe inventory of laptops, servers and services, and who is responsible for eachIT/Operations Lead

Data

Classification, retention, encryption and privacy: the rules that follow the data itself.

PolicyWhat it coversTypical owner
Data Classification and Handling PolicyThe classification levels and the handling rules that follow from each oneSecurity Owner
Data Retention and Disposal PolicyHow long each kind of data is kept, and how it is deleted when the time is upSecurity Owner
Encryption and Key Management PolicyEncryption in transit and at rest, and how keys are managed and rotatedEngineering Lead
Privacy and Data Protection PolicyHow personal data is handled, and the commitments made to the people it belongs toSecurity Owner

Engineering

How software reaches production and how weaknesses are found and fixed.

PolicyWhat it coversTypical owner
Change Management PolicyHow changes to systems are requested, reviewed, approved and releasedEngineering Lead
Secure Software Development PolicySecure development: code review, testing, dependencies and environment separationEngineering Lead
Vulnerability and Patch Management PolicyScanning for vulnerabilities and the deadlines for fixing what is foundEngineering Lead
Logging and Monitoring PolicyWhat gets logged, how long logs are kept, and which events raise an alertEngineering Lead

Resilience

What happens when something breaks, and how you get back.

PolicyWhat it coversTypical owner
Incident Response PolicyHow an incident is declared, who is called, and what is recorded afterwardsSecurity Owner
Business Continuity and Disaster Recovery PolicyRecovery objectives and the plan for continuing to operate through a disruptionEngineering Lead
Backup and Recovery PolicyBackup schedule, retention and encryption, and how restores are testedEngineering Lead

People, devices and suppliers

The controls that live outside the codebase and are the easiest to forget.

PolicyWhat it coversTypical owner
Vendor and Third-Party Risk Management PolicyAssessing vendors before they get data, and reviewing the important ones each yearSecurity Owner
Human Resources Security PolicyScreening, onboarding, training, acknowledgment and offboarding of peoplePeople/HR Lead
Endpoint and Workstation Security PolicyLaptop and workstation controls: encryption, screen lock, updates and managementIT/Operations Lead
Network and Infrastructure Security PolicyNetwork boundaries, firewall rules, segmentation and remote accessEngineering Lead
Physical and Remote Work Security PolicyPhysical security for offices and equipment, and the rules for working remotelyIT/Operations Lead

What separates a usable template from a useless one

The difference is not length or how formal the language sounds. It is whether the document describes something that happens. Five things make that difference in practice:

  • A named owner per policy, meaning a person or a role that exists at your company, not “the Information Security Committee” at a company of eight people.
  • Real cadences. If the policy says access is reviewed quarterly, that review has to happen four times a year and leave a record. Pick the slowest cadence you will genuinely keep.
  • The tools you actually use, named. A policy that says “the identity provider” is vague; one that says Google Workspace, with MFA enforced, is testable.
  • An approval and a date, plus a revision history that gains a row every time the policy is reviewed, even when nothing changed.
  • Acknowledgment records showing that the people bound by the policy have read it.

A template that arrives with bracketed placeholders is doing you a favour by making the gaps visible. The risk is the opposite kind: a polished document full of controls that sound plausible and that nobody at your company performs. That is a written statement contradicted by your own evidence, which is worse than having written nothing.

What to change before you adopt one

Work through the set once with this in mind: replace every role with a person, every frequency with one you will keep, and every generic tool reference with the product you pay for. Delete whole sections that do not apply, for example physical data centre controls if you are entirely on a cloud provider, and say in one line why they do not apply rather than leaving them silently missing.

Then get it approved and acknowledged. An unapproved policy is a draft, and a policy nobody has acknowledged is a document rather than a rule. The acknowledgment guide covers how to collect and keep those records.

If you are heading for SOC 2

The same set is what a SOC 2 examination asks you to produce, which is why each policy on this site is mapped to the Trust Services Criteria it addresses. Two things are worth knowing early: there is no official list of SOC 2 controls, so you write your own and map them to the criteria (see the controls list guide), and policies alone are not enough, because the auditor tests whether the controls described actually operated during the period. For the order in which to adopt them, see SOC 2 policies for startups; for the full index with the criteria mapping, see the template index.

Frequently asked questions

What should an information security policy contain?
A statement of scope and objectives, the roles responsible for security, the rules that apply to everyone, and a reference to the supporting policies that go into detail. It should also name an approver, carry an effective date and say how often it is reviewed. Anything longer than a few pages usually belongs in a supporting policy instead.
How many security policies does a small company need?
Fewer than most template packs imply, but more than one. The set here is 22 documents because separating them by owner and review cadence is easier to maintain than a single long document. A five-person company can adopt all of them and keep several very short.
Are these templates free?
Yes. The 22 templates are Apache-2.0 licensed and you can generate the whole set, edited for your company, without signing up or giving an email address. The paid Audit Kit is optional: it rewrites the policies around the specific tools you use and produces Word documents, a criteria crosswalk and a review calendar.
Can I just copy a security policy template and be done?
Not safely. A policy that describes controls you do not operate is worse than no policy, because it is a written statement that does not match reality. Change the owners, the cadences and the tool names to what you actually do, and delete the sections that do not apply.
Do these templates work for SOC 2?
They are written for it: the set maps to the Trust Services Criteria, and each criterion has a page showing which policies address it and what evidence an auditor typically samples. Policies are only part of an examination, though. The controls they describe have to operate and be evidenced over the period.
Who should own each policy?
Someone who can actually make the decision it describes. In practice that means a security owner for governance, access and data policies, an engineering lead for change, development, logging and infrastructure, and an operations or people lead for devices, physical security and HR. The templates come with a suggested owner per policy, which you change to a real name.

Policyseed provides governance policy templates and AI tailoring. It is not legal advice and not a compliance guarantee. Management adopts the policies; the CPA firm performs the SOC 2 examination.