Introduction

DevWorkWire brings already refined Epic and Story Markdown into Jira through one local CLI, dwire. Use it from your terminal or let a terminal-capable AI agent use the same commands through a portable skill.

Start with the project overview for the problem and scope, or go directly to the CLI contract for supported commands and recovery behavior.

What you can do today

• Fetch individual Epics and Stories, list an Epic's Stories, and list open work assigned to a user.
• Create individual Epics and Stories using direct CLI commands.
• Validate and preview a prepared Markdown folder locally, without Jira writes.
• Import missing items and resume using the local record of created Jira keys.
• Use direct commands with dwire --format json COMMAND ... from scripts and terminal-capable agents.

Run dwire without a command for the interactive menu. For installation and configuration, follow the application repository's CLI reference and AI agent skill setup guide.

From Markdown to Jira

These examples assume dwire is installed and Jira access is configured. Replace ./work-items/my-epic with your prepared folder.

1. Prepare the work. Put one Epic in epic.md and its optional Stories in stories.md. DevWorkWire imports defined work; it does not refine an idea into requirements or acceptance criteria.

2. Preview locally. Inspect validation errors and the items marked for creation, skipping, or resolution before any Jira write.

   dwire preview-folder ./work-items/my-epic

3. Review and import. The import command validates and previews again, then asks for confirmation in an interactive terminal.

   dwire import-folder ./work-items/my-epic

4. Keep the resume file. Preserve .devworkwire-import.json beside your Markdown. Re-running with that file skips recorded items and can continue missing ones. Changed uploaded items are skipped, not updated in Jira.

Write boundaries and recovery

Direct create-epic and create-story commands write immediately. JSON-mode and other non-interactive imports require --yes when there is work to create. Agents must stay within the user's approved task; the CLI cannot verify a conversation's authorization.

Without the resume file, the importer cannot identify earlier uploads from Jira alone. If a write's outcome is uncertain, inspect Jira before resolving or retrying it. Read the write and recovery rules.

Choose a reading path

Your goal	Start here	Continue with
Understand the product	Overview	Personas and scope boundaries
Use or automate the CLI	CLI contract	CLI-first agent integration
Contribute to implementation	Feature map	Current solution design and decisions
Review future direction	Planning	Prototype references

Explore the documentation

Section	What you will find
Context	Overview, personas, terminology, and scope
Requirements	Feature requirements and implementation status
Planning	Draft roadmap, roles, and workstreams
Architecture	Core design, interfaces, persistence, diagrams, operations, and security
Decisions	Accepted, proposed, and superseded Architectural Decision Records
Prototype	Prototype brief, design direction, and interaction references

Current behavior and proposals

Some architecture and requirement pages retain earlier target designs. A Clarified requirement or Proposed ADR is not proof of an implemented feature. Use the feature map, current CLI contract, and accepted ADR-011 to interpret them.

MCP, the SQLite preview store, Jira-side matching, updates, richer queries, comments, status transitions, and PR references remain deferred or planned. The current integration is the installed CLI plus a portable skill.