Jekt for VS Code

The Jekt™ context architecture maintains software project context alongside the project's code. Tickets, plans, designs, minutes, and comments are persisted as Markdown and YAML in a separate Jekt directory tree, so developers and AI assistants can read and edit the same project state. Storage in a Git repository allows this context to persist across agent sessions, ticket assignments, and releases. Jekt requires no separate server or account; its file layout is defined by an open specification.

For example, a ticket to fix portrait photos appearing sideways in thumbnails might contain:

.jekt/tickets/photos-42/
├── description - Sideways Thumbnails.md
├── description.yaml
├── plans/
│   └── 2026-09-25-correct-orientation.md
├── designs/
│   └── Image Orientation.md
├── minutes.md
└── comments/

This ticket's description Markdown document describes the bug, while its metadata YAML file records ticket properties such as status and priority. The plan guides a developer or agent to change thumbnail generation to fix the bug. The design records how orientation is handled as plans converge on a cohesive understanding. The minutes capture decisions made during the fix. These files all concern the same ticket, and allow insight to be surfaced and persisted as development progresses.

Extension Capabilities

This extension provides tools for working with Jekt in VS Code. Its first release includes a Jira Cloud importer so you can start using Jekt with an existing project's work and history already in place. It converts the project's issues and supporting content from a downloaded Jira Cloud backup into Jekt files in your repository. You can continue working with that accumulated context: browse and edit the files, track changes with Git, and use them with AI assistants.

Try the Jira Cloud import functionality without commitment to Jekt: you can always delete the imported data or archive the import as a better-organized backup than the original Jira Cloud backup. Test it in your project, or in a separate, empty folder. The import process reads your downloaded backup without connecting to Jira, changing Jira, or changing the backup. It writes Jekt files under .jekt/ in the destination and does not stage or commit them. If you decide not to keep an import that was only for trial, the added .jekt/ folder can simply be deleted, unless the destination you chose already contained a .jekt/ directory with information you want to keep.

Regardless of the source of the Jekt information, you can use Jekt's reference prompts to direct AI assistants to find and understand project context and work with Jekt artifacts. Those prompts are available separately and are not installed by this extension. Jekt agent skills are on the roadmap.

Import a Jira Cloud Project

The Jekt: Import Jira Cloud Backup... command imports a Jira Cloud project's issues and supporting content into your repository as Jekt files you can work with outside Jira. The import reads a downloaded Jira Cloud backup, which is an exported snapshot of the Jira site's projects and their associated issues and content. Each run imports one Jira project from the backup, creating a Jekt ticket for every Jira issue and preserving each Jira issue key and its original casing as the Jekt ticket ID.

For example, an issue about audio stopping when switching presets in a guitar-effects application might have the Jira key GUITAR-117 and produce:

.jekt/tickets/GUITAR-117/
├── description - Audio Stops When Switching Presets.md
├── description.yaml
├── comments/
└── attachments/

The description document holds the issue's title and description, and description.yaml records its key fields as ticket metadata. Comments and available attachments are saved under comments/ and attachments/, respectively. Jekt's more semantically rich artifacts, such as plans, designs, and minutes, can be added as work proceeds; the import does not create them because Jira does not collect that information. Note that Jekt normally uses lowercase ticket IDs, but imports preserve Jira issue keys and their casing for consistency with Jira.

Import Scope

The import brings in:

Issue titles, descriptions, comments, and available attachments.
Supported fields such as type, priority, status, resolution, and assignee.
The selected project's original Jira key and, when possible, its name and description.

Descriptions and comments are converted to Markdown, and links between issues in the selected project point to their Jekt tickets. Project details are recorded under .jekt/projects/, unless an existing Jekt project is declared there already.

This extension release skips restricted comments and their attachments to avoid exposing content governed by Jira's access controls. The import log reports these exclusions and any attachments that were missing from the backup or could not be written. Future versions may provide more visibility-related options.

In this version, imported tickets are not a complete copy of Jira: fix-version releases, labels, components, sprints, custom fields, and structured relationships such as epics and dependencies are not yet imported. Future versions are expected to expand this coverage through additive imports. These imports add missing content; they do not keep Jira and Jekt synchronized. For specific details about field mappings, read the Jekt Jira Profile.

Prepare a Jira Cloud Backup

Create and download a Backup for cloud using Atlassian's export guide. Ask your Jira administrator if you cannot create one. Including attachments is recommended; imported attachments will be placed inside attachments/ for each Jekt ticket, and references to those attachments in ticket descriptions and comments will be preserved.
Extract the downloaded ZIP file.

Run the Import

Use VS Code 1.125 or later in a trusted workspace. The extension is disabled in Restricted Mode. You do not need GitHub Copilot, a Jira API token, or a Jekt account to import.

Open a destination folder in VS Code.
Run Jekt: Import Jira Cloud Backup... from the Command Palette (Ctrl+Shift+P on Windows/Linux or Cmd+Shift+P on macOS). In a multi-root workspace, select the destination root folder when prompted.
Select the extracted folder containing entities.xml, choose the Jira project, and enter its Jira Cloud domain, such as example.atlassian.net. The domain is used to record each ticket's Jira source in metadata, verify ticket identity on repeat imports from the same domain, and recognize and rewrite links to other tickets within the project. The import process does not connect to Jira.
Review the project, backup, domain, and destination, then select Import.

The import creates the .jekt/ directory if needed and adds its orientation readme if missing, without replacing existing files. This initialization does not include features such as team coordination; you can add them later by initializing again with a Jekt skill or prompt.

Review and Use the Imported Tickets

The imported tickets appear under .jekt/tickets/, which you see in VS Code's Explorer View. Select Show Log when the import finishes to see more details about the import activities, the results, and anything skipped.

Open and search the tickets, review and edit their files, and commit them with Git if you want to keep them. Importing again adds missing content without replacing existing ticket descriptions or field values.

Add a plan, record a decision in minutes, or create a new Jekt ticket alongside the imported ones—directly or using the reference prompts or skills. The Jekt Primer explains these artifacts and how to use them, whether you're starting with a Jira import or from scratch in an empty Jekt repository.

Privacy

This version of the extension does not collect or send telemetry. GlobalMentor receives no information from it, as the extension sends nothing over the network. The extension reads and writes only files in your workspace and in locations you select, whether on your local device or a remote one. VS Code remembers some of your recent entries for later use, such as the last folder you imported from and site addresses you entered, and keeps the extension's log, shown in the Jekt output channel, in its log files.

Data imported from other systems commonly contains personal data, such as people's names, usernames, and email addresses. The extension writes some of it into your repository: it records the people who created, were assigned, commented on, attached files to, or were mentioned in each imported item, by the usernames or other identifiers the source system uses, and it writes the descriptions, comments, and attachments themselves, with whatever their authors put in them. You decide what to import and share. Imported files may contain private project information; review them before committing or sharing them. VS Code, the Marketplace, and any AI assistant you choose to use have their own data practices.

Questions

See the Jekt FAQ for common questions, or ask about the extension on its Marketplace Q & A tab.

License

By installing or using this extension, you agree to the Jekt Extension License Agreement.