Jira story, epic, and bug
When intake.jira.enabled is true and the ask contains an issue key, the receptionist resolves jira-story, jira-epic, or jira-bug from the issue type — before change class.
Shared rules
intake-agentfetches once. Everyone else readsintake.md. Do not call the tracker from the parent or from BA.- The configured
intake.jira.connection(mcp,cli, orapi) is the only fetch path. If it fails, intake returnsBLOCKED. Switchconnection, fix auth, or paste the description as plain text (feature-development). Do not guess. - Do not hardcode a project key, JQL, MCP server name, CLI argv, or issue-type mapping in a skill. Those belong in
config.json.
Chains
| Workflow | Plan source | Chain |
|---|---|---|
jira-story | intake.md | intake → sign-off requirements → architect? → BA → waves → tester → devops → retro (skip_pm) |
jira-epic | epic-plan.md | Same shape; BA writes one child spec per in-scope story |
jira-bug | rca.md | intake → bug-analyst → developer → critic → tester → devops → retro. No PM/BA/Architect |
run_tester is true on bugs. Do not let the developer start a fix before HANDOFF-bug-analyst.md is SUCCESS. rca.md is the defect’s specification.
Issue-type map (story, task, bug, defect, epic, …) is under intake.jira.issue_type_map. For connection: mcp, tool names default to getJiraIssue and searchJiraIssuesUsingJql.
Enable/disable: config guide. Failures: intake troubleshooting.