nightshift:create

Create a new Nightshift shift named $ARGUMENTS.

Steps

1. Resolve the shift name

   • If $ARGUMENTS is non-empty, treat it as the shift name.
   • If $ARGUMENTS is empty, use the AskUserQuestion tool (open-ended) to ask:

     "What will this shift do? Describe the batch work and I'll derive a name."

   • Derive a kebab-case name from their description (e.g., "create promo example pages" → create-promo-examples).

2. Validate the name

   • Must be kebab-case: lowercase letters, numbers, and hyphens only.
   • If invalid, report the error and ask for a valid name.

3. Check for conflicts

   • If .nightshift/<name>/ already exists, report: "Shift <name> already exists. Use /nightshift:start <name> to resume it."
   • STOP if conflict found.

4. Run the bundled init script

   ${CLAUDE_SKILL_DIR}/scripts/init-shift.sh <name>
   

   The script does two things, both idempotent:

   • Project bootstrap (runs every time, no-op when already set up): ensures .nightshift/archive/ exists and creates .nightshift/.gitignore with default ignore patterns if absent. This replaces the work nightshift init used to do at install time — a fresh project never needs a separate "init" step.
   • Shift creation: creates .nightshift/<name>/ containing manager.md with default Shift Configuration and an empty Task Order, plus table.csv empty (no header row — columns are added when tasks are defined).

5. Ask about initial tasks

   Use the AskUserQuestion tool to ask:

   "Shift <name> created! Do you want to add a task now?"

   Options: "Yes, add a task" / "No, I'll add tasks later"

   If yes, guide them through the /nightshift:add-task <name> flow.

Output

## Shift Created: <name>

**Location:** `.nightshift/<name>/`
**Files:** manager.md, table.csv

### Next Steps
- `/nightshift:add-task <name>` — add tasks to the shift
- `/nightshift:update-table <name>` — add items to the table
- `/nightshift:start <name>` — begin execution (after adding tasks and items)

### Environment Variables (Optional)
Create `.nightshift/<name>/.env` to define environment variables for this shift.
Use `{ENV:VAR_NAME}` in task steps to reference them. The `.env` file is gitignored.

Available template variables in task steps:
- `{column_name}` — table item column value
- `{ENV:VAR_NAME}` — environment variable from `.env`
- `{SHIFT:FOLDER}` — shift directory path
- `{SHIFT:NAME}` — shift name
- `{SHIFT:TABLE}` — full path to table.csv

Guardrails

• Always validate kebab-case naming.
• Never overwrite an existing shift.
• Use today's date (ISO YYYY-MM-DD) for the created field — the script handles this.