Jira

If not found, ask the user for both values via AskUserQuestion, then save to memory:

• Write to /Users/larryhsiao/.claude/projects/-Users-larryhsiao-skadi/memory/jira_config.md
• Add pointer to MEMORY.md

Require JIRA_API_TOKEN env var. If not set, tell the user:

Set JIRA_API_TOKEN in your environment (e.g. in ~/.zshrc) with an Atlassian API token from https://id.atlassian.com/manage-profile/security/api-tokens

2. Parse arguments

• type: task, bug, or epic — default to task if not provided or not one of those values
• title: everything after the type (if the second word is not a valid type, treat the entire argument as the title with type defaulting to task)
• Map type to Jira issue type names: task → Task, bug → Bug, epic → Epic

3. Auto-detect Jira project key

Search git history and branches for Jira ticket patterns ([A-Z]+-[0-9]+):

# Try git log first
git log --oneline -100 2>/dev/null | grep -oE '[A-Z]+-[0-9]+' | head -1

# Fall back to branch names
git branch -a 2>/dev/null | grep -oE '[A-Z]+-[0-9]+' | head -1

Extract the project key (e.g. PROJ from PROJ-123) and use it silently — do NOT ask the user to confirm.

If no ticket is found anywhere:

• Check memory file jira_project.md for a saved project key
• If still not found, ask the user for the project key, then save to memory:
  • Write to /Users/larryhsiao/.claude/projects/-Users-larryhsiao-skadi/memory/jira_project.md
  • Add pointer to MEMORY.md

4. Get issue summary

If a title was provided as an argument, use it directly — do NOT ask the user.

Otherwise, use AskUserQuestion to get the issue title/summary. Keep it concise.

5. Fetch sprints and ask user to pick

a. Get the board ID for the project:

curl -s -u "$JIRA_EMAIL:$JIRA_API_TOKEN" \
  "$JIRA_BASE_URL/rest/agile/1.0/board?projectKeyOrId=PROJECT_KEY" \
  | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['values'][0]['id']) if d.get('values') else print('')"

If no board is found, skip sprint selection entirely.

b. Get active and future sprints:

curl -s -u "$JIRA_EMAIL:$JIRA_API_TOKEN" \
  "$JIRA_BASE_URL/rest/agile/1.0/board/BOARD_ID/sprint?state=active,future" \
  | python3 -c "
import sys, json
d = json.load(sys.stdin)
for s in d.get('values', []):
    print(s['id'], s['name'], s.get('state',''))
"

c. Present sprint options via AskUserQuestion:

List each sprint as an option (name + state), plus a "No sprint (backlog)" option.

If the user picks a sprint, record the sprint ID for the issue creation payload.

6. Create the issue

Build the payload and POST to Jira:

python3 -c "
import json, subprocess, os, sys

payload = {
    'fields': {
        'project': {'key': 'PROJECT_KEY'},
        'summary': 'SUMMARY',
        'issuetype': {'name': 'ISSUE_TYPE'}
    }
}

# Add sprint if selected (classic boards use customfield_10020)
# payload['fields']['customfield_10020'] = SPRINT_ID

data = json.dumps(payload)
result = subprocess.run([
    'curl', '-s', '-X', 'POST',
    '-u', f\"{os.environ['JIRA_EMAIL']}:{os.environ['JIRA_API_TOKEN']}\",
    '-H', 'Content-Type: application/json',
    '-d', data,
    f\"{os.environ['JIRA_BASE_URL']}/rest/api/3/issue\"
], capture_output=True, text=True)
print(result.stdout)
"

Use the sprint field customfield_10020 (integer sprint ID) for classic Jira boards. If it returns a field error for that field, retry without the sprint field and note that the user should assign the sprint manually.

For next-gen (team-managed) projects, sprint assignment is done via the agile API after creation:

curl -s -X POST -u "$JIRA_EMAIL:$JIRA_API_TOKEN" \
  -H "Content-Type: application/json" \
  "$JIRA_BASE_URL/rest/agile/1.0/sprint/SPRINT_ID/issue" \
  -d '{"issues": ["ISSUE_KEY"]}'

7. Report the result

Parse the response and show:

Created: PROJECT-123
URL: https://company.atlassian.net/browse/PROJECT-123

Then ask via AskUserQuestion: "Start working on PROJECT-123 now?" (yes → chain to /working PROJECT-123).

Verb: status

Shows a sprint board overview grouped by status, sorted by priority. Acts as a secretary: tells you what the team is doing and what you should focus on next.

Arguments after status: /jira status [filter]

• No argument: show all issues in the current sprint
• me: show only issues assigned to the current user
• Any other string: filter by that person's display name

1. Load Jira config from memory

Check memory file jira_config.md for:

• JIRA_BASE_URL
• JIRA_EMAIL

If not found, ask the user and save to memory (same as create verb).

Require JIRA_API_TOKEN env var. If not set, tell the user to set it and stop.

2. Resolve project key

Same as create verb: git history → jira_project.md memory → ask user.

3. Resolve assignee filter

• No argument → no filter (show all)
• me → use JIRA_EMAIL from memory as the assignee value
• Any other string → use as the assignee display name in JQL

4. Fetch active sprint

a. Get board ID:

curl -s -u "$JIRA_EMAIL:$JIRA_API_TOKEN" \
  "$JIRA_BASE_URL/rest/agile/1.0/board?projectKeyOrId=PROJECT_KEY" \
  | python3 -c "import sys,json; d=json.load(sys.stdin); print(d['values'][0]['id']) if d.get('values') else print('')"

b. Get active sprint (only if board found):

curl -s -u "$JIRA_EMAIL:$JIRA_API_TOKEN" \
  "$JIRA_BASE_URL/rest/agile/1.0/board/BOARD_ID/sprint?state=active" \
  | python3 -c "
import sys,json
d=json.load(sys.stdin)
sprints=d.get('values',[])
if sprints:
    s=sprints[0]
    print(s['name'] + '|' + s.get('startDate','') + '|' + s.get('endDate',''))