Getting Started with Commonset

Commonset gives your organization one place to manage the reusable AI capabilities used across teams and AI platforms.

Instead of skills, instructions, tools, and workflows living independently inside individual provider accounts, Commonset creates a shared system for understanding what exists, who owns it, whether it is trusted, who can use it, and where it is available.

This guide introduces the basic Commonset workflow.

What Commonset Manages

Commonset uses capability as the broad term for a reusable AI function that your organization wants to manage.

A capability might be:

• a Claude Skill
• an OpenAI-compatible capability
• a reusable set of instructions
• a workflow
• a tool configuration
• another provider-specific format

A skill is a specific type or format of capability.

This distinction allows Commonset to manage organizational capabilities independently of any single AI provider.

The Basic Workflow

Most capabilities move through the same general lifecycle:

Discover or create → Review → Approve → Distribute → Use → Measure

Commonset keeps the capability, its versions, ownership, governance state, and provider relationships visible throughout that lifecycle.

1. Add a Capability

Capabilities can enter Commonset in several ways, depending on the integrations and features enabled for your organization.

You may be able to:

• create a capability directly in Commonset
• import an existing capability
• pull a skill from a connected AI provider
• create a new version of an existing capability

When adding a capability, Commonset records information such as its name, source, ownership, version, and provider.

The goal is to establish clear provenance: you should be able to determine what the capability is and where it came from.

2. Review the Capability

Capabilities may be evaluated before they are made broadly available.

Review can include:

• inspecting the capability contents
• comparing the current version with an earlier version
• reviewing ownership and source information
• evaluating automated security checks
• confirming organizational policy requirements
• determining which teams or users should have access

Commonset is designed to provide governance without making experimentation unnecessarily difficult.

Your organization's policies determine when formal review is required and what happens while a capability is being evaluated.

3. Understand Trust Status

A capability version can have a trust status that communicates whether and how it should be used.

Common statuses include:

Draft

The version is still being developed and has not been submitted for organizational approval.

In Review

The version is being evaluated.

Approved

The version meets the applicable trust requirements and can be made available according to your organization's access policies.

Blocked

The version cannot be distributed while the blocked status is active.

Deprecated

The version remains part of the historical record but should generally no longer be used for new work.

Trust status applies to a specific version. A previously approved version does not automatically make a new version approved.

4. Control Access

Approval and access answer different questions.

Approval answers:

Does this version meet our trust requirements?

Access answers:

Who is allowed to use it?

Depending on your organization's configuration, access may be granted to:

• the entire organization
• individual teams
• specific users
• defined groups or policies

This separation allows organizations to approve a capability without automatically making it available to everyone.

5. Publish and Distribute

Once a capability is ready, Commonset can make it available to supported AI platforms.

The organizational capability remains the source of truth while provider-specific copies or representations can be distributed to platforms such as Claude, ChatGPT, internal AI systems, and future providers.

This separation is important.

The provider is where a capability may be consumed or executed.

Commonset is where the organization governs it.

6. Manage Versions

Capabilities change over time.

Commonset treats versions as distinct revisions so teams can understand:

• what changed
• who changed it
• which version is approved
• which version is currently distributed
• whether a newer version requires review

Version history makes it possible to improve capabilities without losing the organizational record of how they evolved.

Where supported, Commonset can also show differences between versions to make review easier.

7. Understand Usage

A capability registry becomes less useful if nobody knows which capabilities are actually being used.

Commonset is designed to connect governance with usage information so organizations can eventually answer questions such as:

• Which capabilities are actively used?
• Which teams use them?
• Which providers are they used with?
• Which approved capabilities have little or no adoption?
• Which capabilities may no longer be necessary?

Availability of usage information depends on the connected provider and your organization's configuration.

A Typical Example

Imagine the Finance team creates an Expense Analyzer capability.

The capability is added to Commonset and assigned an owner.

An initial version is created and evaluated.

After the applicable checks and review are complete, that version becomes Approved.

Access is granted to the Finance team.

Commonset then distributes the approved version to a supported AI provider where Finance employees can use it.

Later, the owner changes the capability.

Instead of silently replacing the approved version, Commonset creates a new version. Reviewers can compare the changes and determine whether the new version should also be approved.

The organization retains a clear record of what changed, what was trusted, and what users were allowed to use.

Commonset's Core Principle

At every important point, Commonset should help you answer:

• What is this?
• Who owns it?
• Where did it come from?
• What changed?
• Is this version trusted?
• Why is it trusted?
• Who can use it?
• Where is it available?

This reflects a core Commonset product principle:

Make trust visible.

Next Steps

Once you understand the basic workflow, continue with:

1. Understanding Capabilities, Skills, and Commonsets
2. Importing Skills from AI Providers
3. Creating and Publishing a Capability
4. Reviewing and Approving Capabilities
5. Managing Access
6. Connecting AI Providers
7. Versioning and Comparing Changes