OpenClaw — Personal AI Automation System
A personal AI workflow system built on Claude Code, MCPs, and custom skills — designed to operate across development, career management, and knowledge work at significantly higher throughput.
- Claude Code
- MCPs
- Obsidian
- Python
- Windows
Problem
Modern AI tools are powerful individually but fragmented — each tool requires its own context, workflow, and mental switching cost. I wanted a unified operating model where AI assistance is persistent, contextual, and integrated into real workflows rather than a separate tool I context-switch into.
What I Built
OpenClaw is a personal system built on top of Claude Code’s plugin and skill architecture:
- Skills: Custom reusable instructions for recurring workflows — career OS operations, code review patterns, content production, session management.
- MCPs (Model Context Protocol servers): Integrations that give the AI direct access to my Obsidian vault, calendar, and file system — enabling it to reason across my actual context rather than isolated prompts.
- Workflow automation: Recurring operations (CV tailoring, interview prep, project documentation, research synthesis) are expressed as skills that invoke the right tools in the right sequence.
- Memory system: Persistent file-based memory across sessions so context accumulates rather than resetting.
Operating Model
I use OpenClaw as a daily operating system for technical work: drafting and maintaining career materials, exploring codebases, designing and implementing features, synthesizing research, and managing the Obsidian vault that backs this portfolio.
Outcome
A measurably faster personal workflow across development, writing, and knowledge management. The system is under continuous iteration — each improvement compounds across all the workflows that depend on it.