Deep Model Trust · R&D
Deep Model Trust · Secure Personal Agent

Bounded autonomy for enterprise agents.

Secure Personal Agent is research into a runtime control layer for tool-using agents: task-specific authority, short-lived credentials, and policy checks before execution.

Research & development · Not commercially available

The proposed approach

Autonomy, minus the blast radius.

This work aims at autonomy inside defined boundaries: task-specific, least-privilege authority, with each proposed tool call evaluated against identity, policy, and context before execution.

01
Scoped

Task-specific, least-privilege authority instead of standing access — read:tickets, not root.

02
Checked

The proposed approach evaluates each tool call against identity, policy, and context before execution.

03
Contained

The design intent is that actions exceeding the approved boundary are blocked, constrained, or escalated, not only logged.

Research in progress.

Secure Personal Agent is part of mace AI’s Deep Model Trust research and development. It is not commercially available.

Explore the research →   Assess your current agent with Fisher