Vestige is the security kernel for AI agents. Your agent can think anything. Vestige decides what it is allowed to do. It remembers every project decision, reaches backward through time to find the quiet choice behind today's failure, and gates dangerous effects: receipt, one-use permit, effect, or STOP with zero effects. Local, encrypted, no cloud, no telemetry. Launching tonight: Founding Operator at $49 for the first month and Managed Continuity at $19/mo.
I built Vestige because I kept watching agents do exactly what they were told, exactly when they shouldn't. The kernel makes stop real: receipt, one-use permit, effect, or STOP with zero effects. The launch video is one take: an agent tries to erase its own memory trail, you watch the refusal land, and a human approves before anything runs. Local-first, free forever, two commands to install. Operator and Continuity went live at midnight. Ask me anything.