Creating an issue¶

Creating an issue takes seconds. The only things Specivo insists on are a subject and a tracker — everything else is optional and easy to add later.

Steps¶

1. Open the project you're working in, then click New issue.
2. Choose a tracker — Bug, Feature, Task, or Support. This sets the issue's type.
3. Write a clear subject. Keep it short and specific: "Login button does nothing on mobile" beats "login broken". The subject can be up to 1024 characters, but shorter is better.
4. Describe the work in the description box. It's Markdown, so you can use headings, lists, tables, and fenced code blocks. Mention a teammate with @username to pull them in.
5. Set the optional fields that apply — assignee, priority, target version, sprint, start and due dates, and an estimate. You can skip any of these and fill them in once the issue exists. If you are allowed to set issues private, there is also a Private box. It starts ticked for a tracker that makes issues private by default (see below).
6. Click Create (or Save).

That's it. The issue gets the next number in the project (for example ACME-57) and opens on its own page.

Similar issues¶

While you type the subject of a new issue, Specivo looks for issues in the same project whose subject is close to it and lists up to five of them under the subject box, the closest first. It asks once you stop typing for a second and the subject has at least 10 characters (4 in Chinese, Japanese, Thai and the other scripts written without spaces). Each row shows the issue's key, its subject and its status, closed issues included; the link opens the issue in a new tab, so your draft stays as it is. Only issues you are allowed to see are listed.

If the issue you are about to file already exists, click Mark as duplicate on its row (shown when you may manage issue relations). It is added to the relations below the form as "Duplicates" and is created together with the issue; Remove takes it off again. You can still create the issue either way: the list is only a hint.

The suggestions compare the words and the spelling of subjects, so they work without any AI or embedding model installed. They do not read descriptions. An administrator can turn them off for the whole instance under Administration → Settings → Defaults, and a project manager for one project in the project's settings.

For titles in a script written without spaces (Chinese, Japanese, Thai, Lao, Myanmar, Khmer), the newest 2000 issues of the project are compared, so an older duplicate is not found. These suggestions also need a database whose character type (lc_ctype) is not C; on such a database they list nothing. The bundled database is set up correctly.

Required vs optional¶

Private by default¶

An administrator can mark a tracker Private by default under Administration → Trackers — a Security or HR tracker, say. Every new issue of that tracker is then created private, from the web form, the API, an AI agent or a recurring task, without anyone having to remember to tick a box. Only members with the Set issues private permission can untick it; everyone else gets the tracker's default. The same permission is needed to make an issue of an ordinary tracker private.

A private issue is visible only to its author, its assignee and administrators. It stays out of everything else too, for everyone who may not see it: issue lists, board and backlog columns, version and sprint pages, counts and progress bars, recurring-task pages, the dashboard, notifications and emails, and issue-key links in text. A hidden key written in a comment looks exactly like a key that does not exist.

After creating¶

• Add structured fields like story points or a git branch — see Issue metadata.
• Link it to related work — see Relations.
• Break large work into subtasks.
• As work progresses, update the status and % done.