issuelisted

# Issue → Plan Turn a ticket into an implementation plan without manual context switching. ## Step 1: Parse arguments `$ARGUMENTS` can be: - GitHub issue number: `45`, `GH-45`, `#45` - Jira ticket: `SOC-123`, `PROJ-456` - URL: `https://github.com/owner/repo/issues/45` - Empty → ask: "What's the issue number or ticket?" Normalize to a clean identifier for fetching. ## Step 2: Fetch the issue **GitHub (try first):** ```bash gh issue view <number> --json title,body,labels,assignees,milestone,comments ``` If that succeeds → extract: - `title` — the issue title - `body` — full description, acceptance criteria, reproduction steps - `labels` — bug / feature / chore / security etc. - `comments` — any clarifying discussion (especially last 3–5 comments) If `gh` fails (not a GitHub repo, not authenticated, or it's a Jira ticket): → Ask: "I couldn't fetch the issue automatically. Please paste the issue title and description." → Continue with the pasted content. **Jira:** ```bash # If JIRA_API_TOKEN and JIRA_BASE_URL are set: curl -s -H "Authorization: Bearer $JIRA_API_TOKEN" \ "$JIRA_BASE_URL/rest/api/3/issue/<ticket>" | jq '{ title: .fields.summary, body: .fields.description, labels: [.fields.labels[]], status: .fields.status.name }' ``` If env vars not set → fall back to paste. ## Step 3: Understand the issue From the fetched content, extract: - **What problem does this solve?** (not just what to build) - **Acceptance criteria** — explicit or implied fro