Ticket Intake

State file: .claude/swe-state/{ticket-id}.json

Fetch a ticket, parse requirements, present a summary for user confirmation.

Usage

/ticket_intake PROJ-123
/ticket_intake https://myorg.atlassian.net/browse/PROJ-123
/ticket_intake #42

---

Phase 1: Detect System

Input Pattern	System	Fetch Method
URL with atlassian.net or jira	Jira	MCP: getJiraIssue
URL with linear.app	Linear	MCP: linear_get_issue
URL with github.com, or bare #N	GitHub	gh issue view
PROJ-123 pattern	Jira (default)	MCP: getJiraIssue
Unrecognized	Unknown	Prompt user

Note: Tool names are for the official Atlassian Rovo and Linear MCP servers. Community servers may use different names (e.g., jira_get_issue for sooperset/mcp-atlassian).

Extract ticket ID from the input (Jira key from path, GitHub number, Linear ID).

---

Phase 2: Fetch Ticket Data

Jira: getJiraIssue with issueIdOrKey. If MCP unavailable:

claude mcp add --transport sse atlassian https://mcp.atlassian.com/v1/sse

Linear: linear_get_issue with issueId. If MCP unavailable:

claude mcp add --transport sse linear-server https://mcp.linear.app/sse

GitHub: gh issue view <N> --json title,body,comments,labels,assignees,milestone

Generic: Ask user for a curl command or API endpoint.

---

Phase 3: Extract Fields

Parse fetched data into structured fields: title, description, acceptance criteria, comments (with author/timestamp), linked issues, labels, priority, assignee.

For acceptance criteria, search the description for:

• - [ ] / - [x] checkboxes
• Numbered lists after "Acceptance Criteria", "AC", "Requirements", or "Definition of Done" headings
• Bullet lists under "Expected Behavior" or "Success Criteria"

---

Phase 4: Identify Affected Areas

Extract keywords (file names, function names, endpoints, class names) from the ticket. Use Grep and Glob to search the codebase. Categorize matches as existing (needs modification), new (referenced but doesn't exist), or test (related test file). Deduplicate and rank by relevance.

---

Phase 5: Present Summary & Write Ticket File

Display the summary (source, title, priority, requirements, acceptance criteria, linked issues, affected areas). Ask the user to review — they can edit, add context, or confirm.

On confirmation, write .claude/swe-state/{ticket-id}/ticket.json:

{
  "ticket_id": "",
  "source": "jira|linear|github|generic",
  "url": "",
  "title": "",
  "description": "",
  "acceptance_criteria": [],
  "comments": [{"author": "", "body": "", "timestamp": ""}],
  "linked_issues": [{"id": "", "relationship": "", "title": ""}],
  "labels": [],
  "priority": "",
  "assignee": "",
  "affected_areas": [{"path": "", "category": "existing|new|test", "reason": ""}]
}

Then update pipeline state: set stages.intake.completed = true.