| name | cb-jira |
| description | Jira Cloud conventions for Cloudbreak — project keys, Team field (customfield_10001), JQL with Atlassian team ids, mascot clan → module mapping, quarterly epic labels. |
Cloudbreak Jira playbook
Use this playbook when answering Jira questions, writing JQL, or using acli, the Atlassian MCP, or the REST API against CB. Do not assume it is loaded for generic coding tasks; open .agent/skills/cb-jira/SKILL.md when Jira context is needed.
Access (tooling & auth)
Two ways to talk to Jira Cloud (https://cloudera.atlassian.net):
acli (Atlassian CLI) — preferred when available. acli jira workitem create|edit|assign|search|comment|view. Authenticate with acli jira auth (OAuth), or with an API token exported as JIRA_EMAIL + JIRA_TOKEN environment variables.- Atlassian MCP (via the Docker MCP gateway) — alternative when
acli isn't set up.
acli gotchas in CB:
- Team (
customfield_10001) can only be set at create — pass it in --from-json under additionalAttributes (e.g. {"customfield_10001": "<team-id>"}); workitem edit rejects additionalAttributes. Verify it stuck with JQL (below), not workitem view (which under-reports custom fields).
- Assign to a real person, not the OAuth principal.
--assignee @me can fail with "cannot be assigned issues" because the OAuth identity isn't an assignable CB user; assign by the person's account id or email instead: acli jira workitem assign --key CB-XXXXX --assignee <email|accountId>.
Project
- Default project keys:
CB or CLOUDBREAK (prefer CB in JQL unless you specifically need the other).
Team field (native Jira Cloud Teams)
- Custom field id:
customfield_10001
- JQL clauses:
Team or cf[10001]
- Display names:
{MASCOT} Clan (e.g. Raven Clan, Wolf Clan, Bear Clan).
JQL: use Atlassian team id, not display name
Clauses such as Team = "Raven Clan" typically return no rows. Filter with the Atlassian team id from issue JSON, the Jira UI, or your integration.
Example (Raven Clan):
project = CB AND Team = 575d3206-8e71-4f04-9e7e-18bb73a1a118-283 ORDER BY updated DESC
Resolve other clans’ ids from customfield_10001 on a representative issue (REST/MCP) or from Jira, depending on your permissions.
Domain mapping (Mascot Clan → repository)
Same mapping as .agent/WORKFLOW.md:
| Clan | Primary areas |
|---|
| Raven Clan | core/, freeipa/ |
| Wolf Clan | datalake/, environment/ |
| Duck Clan | redbeams/ |
| Bear Clan | autoscale/ |
| Hawk Clan | cloud-aws-*, cloud-azure/, cloud-gcp/ |
Labels and epics
- Quarterly epics: Labels such as
Q4FY26-CB.
Confluence (CB wiki scope)
CB engineering docs live under the Cloudbreak Engineering Home hub (page id 92898027, space ENG). Scope searches to that subtree with the ancestor filter rather than the whole space, e.g.:
type = page AND space = "ENG" AND ancestor = 92898027 AND text ~ "your keyword"
Pull keywords from the Jira summary, error text, labels, or components. When search returns a page URL, fetch by the numeric page id (the segment in /wiki/spaces/.../pages/<pageId>/...) to read the full body. If the ticket links outside this subtree, still fetch those pages — the ancestor filter is only the default discovery scope.
Commits and PRs
- Commit messages use
CB-XXXXX Subject so CI can link PRs to Jira (see root AGENTS.md — Operational mandates and Continuous integration).
