shared/input-classifiers/jira-url
Source:
shared/input-classifiers/jira-url.md
input classifier: Jira URL
Pattern
https?://<site>.atlassian.net/browse/<KEY>-<NUM>https?://<site>.atlassian.net/jira/.../browse/<KEY>-<NUM>Fetch via
adk-mcp-atlassian → getJiraIssue(cloudId, issueIdOrKey) (the mcp-atlassian tool exposes the call directly).
Fallback (if MCP unreachable): direct REST.
curl -sS -u "$ATLASSIAN_USERNAME:$ATLASSIAN_API_TOKEN_CRED" \ "https://$ATLASSIAN_SITE/rest/api/3/issue/$KEY" \ | jq '{key, summary: .fields.summary, status: .fields.status.name, type: .fields.issuetype.name, assignee: .fields.assignee.displayName, description: .fields.description, sprint: .fields.customfield_10020}'Extract into context.md
### [jira] <KEY> — <summary>url: <full URL>status: <In Progress | Done | …>type: <Story | Bug | Spike | …>assignee: <name>sprint: <name if any>summary: <one paragraph from description, ≤80 words>acceptance criteria: - <bullet> - <bullet>linked issues: - blocks: <KEY> - blocked-by: <KEY>attachments: <count, names>key comments: <last 3 if substantive>Hints for downstream skills
/adk-implement: acceptance criteria → success criteria in plan.md./adk-review: if linked PR exists, fetch it./adk-investigate: if linked incident/RCA, fetch./adk-document --type pr-body: cite the ticket key as the trailer.