How I Automated CS, Bug Fixes, and Competitor Monitoring with Claude Code Schedule

#flutter #supabase #claudecode #automation

The Problem

I run a personal SaaS called "Jibun Corp" (自分株式会社) — an AI-integrated life management platform built with Flutter Web + Supabase. As a solo developer, I was drowning in:

Support ticket responses
Bug reports and fixes
Competitor monitoring across 14 products
Daily reports and metrics tracking
Blog post writing

The Solution: Claude Code Schedule

Claude Code recently added a "Schedule" feature. It runs tasks on a cron-like schedule, entirely in Anthropic's cloud. No server, no PC required, no API costs beyond the Pro plan.

My 9 Automated Tasks

Task	Frequency	What it does
daily-report	Daily 9 AM	Fetches metrics → generates report → posts to X
cs-check	Hourly	Reads tickets → replies via FAQ or fixes bugs → escalates complex ones
weekly-sns-draft	Weekly Mon	Summarizes week → drafts social posts
daily-development	Daily 10 AM	Pushes roadmap tasks forward
pr-auto-review	Every 3h	Reviews open PRs for security, performance, logic
competitor-monitoring	Daily 7 AM	Checks 14 competitors' websites and news
infra-health-check	Every 30 min	Verifies DB connectivity and table availability
dependency-audit	Weekly Mon	Checks for vulnerable packages
blog-draft	Daily 8 AM	Generates technical blog drafts

Architecture

CLAUDE.md (task definitions)
    ↓
Claude Code Schedule (cron trigger)
    ↓
WebFetch → Supabase Edge Functions (thin API layer)
    ↓
Supabase PostgreSQL (data persistence)
    ↓
GitHub push → Firebase auto-deploy

Key Learnings

Edge Functions as thin APIs: Schedule can only use WebFetch (HTTP), so create minimal Edge Functions that expose your data.
RLS matters: Schedule runs as service_role, so set up RLS policies accordingly.
Log everything: Created a schedule_task_runs table to track execution status, viewable from the admin dashboard.