AI agents run locally but often have unclear permissions. Users don't know what files or system resources these agents can access.
Build a lightweight monitor that lists all AI agents on a machine and their access levels. Highlight sensitive areas like documents, cameras, or microphones.
Privacy-focused individuals and companies would pay for this to prevent data leaks. IT departments could use it for compliance.
Start with a simple CLI tool that scans and reports. Later add GUI and real-time monitoring.
The biggest risk is false positives scaring users about harmless agents.