Demo Todoist account with representative data

[craigmccaskill]

Problem

Several features need a Todoist account with known, representative data:

• Demo GIF (Demo GIF: tell a workday story, not a feature showcase #191) — needs realistic tasks across projects to tell a workday story
• Performance benchmarking (CI startup time benchmark with PR reporting #118) — API call latency requires real responses, can't measure with mocks
• Integration testing — if we ever want to test beyond mocks, we need a real account
• Docs screenshots — workflow guides (Docs site: add workflow guides alongside command reference #190) benefit from real output examples

Currently there's no shared test account, so each of these items independently solves the "where do I get realistic data?" problem.

Proposed Solution

Create a dedicated Todoist account with representative data:

• Multiple projects (Work, Personal, Side Project, etc.)
• Tasks spanning priorities, due dates, labels, recurring tasks, overdue items
• Sections within projects
• A mix of inbox items and organized tasks
• Enough volume to be realistic (50-100 tasks)

Setup

• Free Todoist account with a dedicated email
• API token stored as a GitHub Actions secret (TODOIST_DEMO_TOKEN)
• A seed script (scripts/seed-demo-data.py) that resets and populates the account to a known state
• Run the seed script before demo recording or benchmark runs

Design Notes

• The seed script is idempotent — safe to run repeatedly
• Task content should feel real and varied (not "Test Task 1", "Test Task 2")
• Data should exercise edge cases: overdue tasks, no-date tasks, long content, unicode, multiple labels

Acceptance Criteria

• Demo account exists with API token in GitHub secrets
• Seed script creates representative data set
• Seed script is idempotent (reset + populate)
• Documented in CONTRIBUTING.md for maintainer use

Blocks: #191 (demo GIF), #118 (API benchmarking)