Developer to Debug & Stabilize AI Agent (OpenClaw/WSL2)

I run an AI agent (OpenClaw, self-hosted on WSL2/Windows, Gemini model) that handles business admin tasks including SMS (Sinch/MessageMedia API), Gmail, and Google Calendar integrations. The system has developed serious reliability problems and needs debugging and stabilization work. Known issues include: a large script directory (60+ files) with many obsolete "ghost" scripts that need auditing and cleanup; hardcoded API credentials in individual scripts that should be centralized/read from a config file instead; an internal cron/scheduling service that isn't visible via standard crontab and needs to be located and audited; and a timezone handling bug where scheduled jobs used UTC instead of local time, causing messages to send outside allowed hours. Looking for someone comfortable with Python, Linux/WSL environments, cron/scheduling systems, API credential management, and debugging agent/LLM-based automation systems. Deliverables: full script audit and cleanup, centralized and secure credential handling, a documented and verified scheduling setup, and a written report of what was found and fixed.