I’m a UK-based developer building runtime visibility and security tooling for AI-native computing.
I’m currently building CoworkGuard — a local-first runtime monitoring layer for macOS that helps users understand what AI-powered apps, browser tools, extensions, and local agents are doing on their machine.
CoworkGuard focuses on:
- AI app activity monitoring
- outbound connection visibility
- browser AI tooling
- extension and agent behaviour
- local AI runtime observability
- sensitive data flow detection
- behavioural correlation timelines
The goal is simple:
What are your AI tools doing?
As AI assistants, coding agents, MCP tools, and browser AI features gain deeper access to local systems, I believe users need better visibility into:
- what accessed sensitive data,
- which process initiated a connection,
- and what left the machine.
CoworkGuard is designed as a calm, local-first observability layer for that new environment.
You can explore the project here:
→ https://github.com/Katherine-Holland/ClaudeCoworkGuard
Alongside CoworkGuard, I work across AI tooling, developer platforms, runtime monitoring, and AI security concepts focused on transparency, privacy, and human-readable security experiences.
Feel free to explore the repositories below.