Issue Summary

Atlassian products, such as Jira and Confluence, manage vast amounts of data, including thousands of projects, documents, and work items. A single Jira instance can contain over 10,000 work items, and Confluence is similarly expansive.

When users interact with Rovo, they may issue commands to the Chat or Agent that cause the underlying Large Language Model (LLM) to exceed its context window. This can lead to incomplete responses or outputs that do not align with the user's query.

Suggestion

Enhance Rovo to prevent breaches of the LLM context window, ensuring a seamless and accurate experience for users.