AgentShield

Policies

Set the rules. Enforce them automatically.

Active Policies

7enforced

Total Violations

29630d

Compliance Rate

96%of users

Enforcement Types

5 Block1 Warn1 Monitor
Policy NameTypeScopeEnforcementStatusViolations (30d)ComplianceActions
US-Only AI Providers

Block model calls to providers based outside the United States

Provider RestrictionGlobalBlockActive
233 active
93%
Approved Tools Only

Only Cursor, GitHub Copilot, and Claude Code are permitted

Tool AllowlistGlobalBlockActive
453 active
87%
Warn on Infrastructure Cmds

Warn when agents run kubectl, terraform, ssh, or docker push

Command WarningGlobalWarnActive
342 active
90%
Block DeepSeek Models

Block all models from the DeepSeek provider

Model BlocklistGlobalBlockActive
121 active
96%
No AI on PCI Repos

Block AI coding tools on repositories tagged PCI-Sensitive

Repo RestrictionTag: PCIBlockActive
81 active
98%
Block External Data Exfil

Block curl/wget to non-allowlisted domains and raw IP addresses

URL RestrictionGlobalBlockActive
183 active
95%
Monitor Paste Site Access

Log when agents fetch from paste sites (pastebin, gist, etc.)

URL RestrictionGlobalMonitorActive
1561 active
85%

Policy Templates

Quickly create policies from proven presets

Block Non-US Providers

Ensure all AI traffic stays within US-based providers

Use Template

PCI Repository Protection

Block AI tools on PCI-sensitive repositories

Use Template

Command Execution Controls

Warn or block risky shell commands from AI agents

Use Template