superagent_
Agent rules editor with an example destructive command blocked before execution.

runtime guardrails

Set limits for agents

Superagent monitors agent activity on workstations and cloud hosts. It evaluates your rules locally, reports matches, and blocks supported actions when enforcement is enabled.

how it works

  • Local monitoring

    Superagent evaluates command, file, and tool events against local security rules. Findings appear in Superagent and can reach your alert workflow.

  • Action blocking

    Superagent blocks supported commands, file actions, or tool calls when they match an enforced rule.

  • Rules across hosts

    Superagent distributes the rule catalog assigned to each client group. Each endpoint validates an update before activating it.

use cases

Sensitive arguments in Pi commands

Superagent evaluates the local rule and blocks supported Pi commands whose arguments match the personal or health data patterns. It sends redacted alerts to the monitoring workflow.

See the Pi rule example →

Environment files on agent hosts

Superagent evaluates attempts to read environment files against your local rules. Supported hooks can block matching actions, and findings appear in Superagent.

Read about runtime rules →

FAQ

read the docs →

more reading

Set up your first agent host

Pair a client so Superagent can apply local rules and report risky activity.