Skip to main content
The Jira Tool allows DinoAI to work with your Jira tickets — reading task requirements and issue details, and creating, updating, and commenting on issues directly from your development workflow. This tool connects development work to Jira, enabling DinoAI to understand requirements, resolve tickets, implement features, and keep issue status and discussion up to date as work progresses.
Requires Jira integration. Before using this tool, make sure you’ve connected your Jira account in Paradime. See Connecting Jira to Paradime for setup instructions.

​
Capabilities

The Jira Tool gives DinoAI both read and write access to your Jira issues. Specifically, it can:
  • Fetch issue details using issue keys or URLs, extracting task requirements and acceptance criteria
  • List and search issues filtered by project, assignee, status, label — or a raw JQL query
  • Create new issues in a project, with summary, description, issue type, priority, and labels
  • Add watchers to an issue when an agent creates it, or later to an existing issue (agents only, see Adding watchers)
  • Link two issues, for example marking one as blocked by another (agents only, see Linking issues)
  • Update existing issues — change the summary, description, priority, assignee, or labels, and transition issue status (e.g. move a ticket to “In Progress” or “Done”) by naming the target status
  • Add comments to issues, e.g. to document progress or record what was implemented
  • Attach a file shared in Slack to a new or existing issue, so a spreadsheet of column specs travels with the ticket it belongs to
Every issue DinoAI creates is automatically tagged with a dinoai label and includes a “Created by DinoAI” attribution line in its description, so tickets filed by the agent are always identifiable.

​
Using the Jira Tool

  1. Open DinoAI in the right panel of the Code IDE
  2. Reference the Jira issue key or URL you want to work with (e.g., FEAT-123 or a full Jira URL), or describe the issue you want DinoAI to find or create
  3. Add your prompt describing what you want DinoAI to do — read the ticket, implement it, update its status, file a new issue, or leave a comment
  4. Approve Jira actions according to your approval mode: Manual asks before each one, Auto has a reviewer model approve safe actions and ask you about risky ones, and Full auto runs them without asking
  5. Review and implement DinoAI’s suggested actions

​
Example Use Cases

Implementing a Feature from a Ticket Prompt
Result: DinoAI extracts the requirements from the ticket, suggests an implementation approach, and helps create the necessary models or code changes to satisfy the acceptance criteria. Resolving a Production Issue Prompt
Result: DinoAI accesses the ticket details, analyzes the error information, and guides you through creating a fix branch, modifying the necessary files, and testing the solution. Filing a Bug You Just Found Prompt
Result: DinoAI creates a new issue in the DATA project with a clear summary and description of the bug, sets the issue type and priority, and returns the new issue key and a link to it. Closing the Loop After Finishing Work Prompt
Result: DinoAI posts a comment on FEAT-456 describing the implemented changes, then transitions the issue to the “Done” status. Reviewing Your Open Work Prompt
Result: DinoAI lists the open issues assigned to you in the DATA project with their status and priority, so you can choose the next piece of work — and start implementing it in the same session.

​
Adding watchers

Agents can add watchers to Jira issues, so the right people follow an issue from the moment it’s created. create_jira_issue accepts a list of watchers, and add_jira_watchers adds watchers to an issue that already exists. Name each watcher in one of these ways:
  • An exact email address or display name
  • accountId:<id> on Jira Cloud, for example when a user’s email is hidden
  • username:<name> on Jira Data Center
Inactive users, and names that match more than one Jira user, aren’t added. If a watcher can’t be added, the issue is still created and the agent reports which watchers failed, so it can retry with add_jira_watchers without creating a second issue. To have an agent add the same watchers to every issue it creates, write who to add in the agent’s Backstory in the agent builder, for example:
Watchers are available to agents only. DinoAI in the Code IDE can’t add them.

​
Linking issues

Agents can link two existing Jira issues with link_jira_issues. Ask in plain words, such as “Mark DATA-12 as blocked by DATA-9”. Use any link type set up in your Jira site, by either of its phrases (blocks or is blocked by) or by its name (Blocks). The default is relates to. If the type doesn’t exist in your site, the agent gets the list of available types and can try again. Both issues must already exist, and an issue can’t be linked to itself. Linking works with Jira Cloud and Jira Data Center, and is available to agents only.

​
Reading a whole issue

Agents read a Jira issue in a compact view by default: status, assignee, priority, labels, components, fix and affected versions, environment, dates, custom fields by name, attachment names and sizes, and the 20 most recent comments. Values longer than 500 characters are shortened. When an agent needs what was shortened, such as an older comment, a whole description, or a custom field ID to update, it calls get_jira_issue with full set to true. That returns the same readable view with nothing cut, not raw Jira data. A very long issue is saved to a file and the agent reads a preview. The full view is available to agents only.

​
Attaching a Slack file to an issue

When an agent is working in Slack, a file someone shared in the conversation can be attached to a Jira issue with attach_slack_file_to_jira_issue. Ask it to create a ticket and attach the file, or to attach the file to an issue that already exists, including on a later turn in the same thread. The original file is uploaded, not the markdown the model read to understand it, so the spreadsheet on the ticket opens as a spreadsheet.
The agent can only attach files shared in that conversation’s own channel, public, private, or DM. A prompt naming a file from somewhere else is refused, so the tool cannot be used to move a file the bot saw elsewhere into Jira.
Agents that restrict their tools with an allowlist need add_jira_watchers, link_jira_issues, and attach_slack_file_to_jira_issue added in the builder. Agents with a denylist, or with no tool restrictions, get all three automatically.

​
Working with Other Tools

The Jira Tool works well alongside DinoAI’s other capabilities to support your full development workflow:
  • Combine with the Terminal Tool to create branches using Jira-based naming conventions (e.g., feature/JIRA-123) and reference ticket IDs in commit messages
  • Combine with the Confluence Tool to cross-reference ticket requirements against linked specification pages
  • Use alongside Git Lite to commit implementation work that traces back to a specific Jira issue — then have DinoAI comment on the ticket and transition it once the work is merged
You’ll need appropriate Jira permissions for the actions you ask DinoAI to perform: creating, updating, or commenting on issues requires the corresponding permissions in your Jira project. In Manual approval mode, DinoAI asks for your approval each time it needs to read from or write to Jira.

​
Best Practices

  • Use exact ticket references — Include the precise Jira issue key or URL in your prompt to ensure DinoAI accesses the correct ticket
  • Ensure ticket quality — DinoAI works best with well-documented tickets that include clear descriptions, acceptance criteria, and relevant context
  • Verify suggestions — Always review DinoAI’s interpretation of requirements before implementing, especially for complex or ambiguous tickets
  • Review writes before approving — When DinoAI creates, updates, or comments on an issue, check the proposed content in the approval prompt; nothing is written to Jira until you approve it
  • Use status names, not transition IDs — To move a ticket, just name the target status (e.g. “In Progress”, “Done”); DinoAI finds and executes the matching workflow transition automatically
  • Combine with the Terminal Tool — Let DinoAI help with the full workflow from implementation to testing, using the ticket as the source of truth throughout