99-second demo: controlled terminal + browser agent execution via MCP with approval boundaries

ImaginaryMachine9110 · reddit · 2026-09-29

An indie project demos letting an AI agent do real work across terminal and browser via an MCP layer without blanket machine access. A 99-second video shows a cross-surface workflow where the MCP layer separates capability from permission: policy checks before sensitive actions, controlled local operations instead of unrestricted access, and observable/auditable execution. The author is soliciting community input on the minimum approval boundaries before letting agents operate real machines.

Original post →

More from coding & agent

coding & agent channel →