Tickets¶
A ticket is the permanent record for one promised result. It keeps the same identity when the worker, stage, priority, or owner changes.
Why tickets exist¶
A chat message or terminal session can disappear. A ticket does not depend on either one. It keeps the title, project, source, priority, current owner, proof, workflow, and ordered history together.
This lets a new person answer simple questions without reconstructing an old session:
- What result was promised?
- Who is accountable now?
- What step is active?
- What proof supports completion?
- What changed, and in what order?
The parts you will see¶
- Priority states urgency. The available values are
P0,P1, andP2. - Source states where the work came from.
- Custodian is the one person or agent accountable for the ticket now.
- Assignee is a person or agent doing a specific part of the work. This can differ from the custodian.
- Version is a number that changes after an accepted update. Ctower uses it to refuse an update based on an old copy of the ticket.
- Timeline is the ordered history. Old events are not rewritten.
How a ticket changes¶
A new ticket starts open. It can be admitted into active work, deferred until a later review time, blocked, unblocked, resolved, closed, or reopened. Reopening starts a new lifecycle episode. A lifecycle episode is one open-to-close period on the same ticket.
The workflow stage and the Board lane are separate facts. A blocked ticket keeps its stage. Changing an assignee does not change the ticket identity or erase proof.
How to use a ticket¶
On the local shadow browser, open /board and select a card. The ticket page shows the record that the
browser can read. The browser does not create or edit tickets.
From the command line, use:
ctl ticket query <ticket-id> --project-key <project>
ctl ticket timeline <ticket-id> --project-key <project>
ctl ticket assignments <ticket-id> --project-key <project>
Use ticket create or ticket capture to create a ticket when you have an authorized command-line
credential. Writes may be queued when off-host acknowledgement is still pending. Do not repeat a write
with a new command identity. Drain the local spool instead.
See the CLI reference for exact flags.