---
title: "A shared context library for your team and its agents"
description: "Keep the skills, documents and records your team relies on in one place, so every teammate and every agent starts from the same context."
canonical: "https://comment.io/shared-working-memory-for-agents"
last-updated: 2026-10-15
---

# A shared context library for your team and its agents

Keep the skills, documents and records your team relies on in one place, so every teammate and every agent starts from the same context.

## The next task starts with what you do today

The plan you refine, the question you answer, the project record you update: all of it can help with what comes next. Comment keeps that company context in a workspace your team and connected agents can use, including when you change agent providers.

## What is shared working memory?

Shared working memory is the set of documents people and agents use to keep track of the work: what you’re trying to do, what you know, what you’ve decided, and what still needs attention.

In Comment, you can read and change those documents together. When an agent starts a new task, point it at the relevant files so it can pick up from the work you’ve already done.

## Write down how your team does it

A skill is a short document of instructions for a task your team repeats, such as how to write a customer update or triage a bug. Comment tells every connected agent about the workspace’s skills and asks it to read the current version when a task matches. Anyone on the team can improve a skill in the editor.

## Start with the work in front of you

You don’t need to prepare a company-wide library to begin. Draft the brief. Review the proposal. Work through the plan. Keep the useful context with it as you go.

## Put your judgment where it can help again

When a draft gets something wrong, discuss the passage and update the document. If the correction belongs in a shared reference too, update that file as well. The next person or agent who reads it can start with the correction in place.

## Structure the parts you want to build with

Some knowledge is useful as both a document and a record: a project with a status and an owner, or a customer with a next step. Doc DB keeps defined fields and written context in the same document.

Your agent can query those fields. A workspace HTML app or dashboard can show a live view of the records. Your team can return to the source documents to review or update the information behind that view.

## Keep context across providers

Choose the agent that fits the next task. When its connection is supported, you can give it access to the same company documents and records. That shared context lives in Comment, where your people can keep using and improving it as your choice of agent providers changes.

## Give another agent somewhere to begin

At the end of a task, record what is settled, what remains open, and where the supporting work lives. Ask the next agent to read that record before carrying on.

Each agent can connect under its own name and work with the same documents and discussions. You can bring different agents to different parts of the work.

## A launch plan taking shape

You ask an agent to research a launch question and put the findings in Comment. You read the brief, challenge an assumption, and settle the audience with a teammate. The agent updates the plan with that decision and the reasons behind it.

When you’re ready to draft the announcement, point your writing agent at the plan. It has a place to find the decision you want the draft to follow.

## Keep the useful parts current

Keep shared context current as the work changes. Point each agent at the decisions and files it needs for the next task.

## Make the next task easier to start

[Join the waitlist](https://comment.io/)

- [What teams use Comment for](https://comment.io/use-cases)
- [Agent-native editing](https://comment.io/what-is-agent-native-editing)
- [FAQ](https://comment.io/faq)