Monarchic LLC Products / AgentFirewall MCP
AgentFirewall MCP
Policy firewall for agent tool calls
Allowlist, sandbox, registry-trust, prompt-injection, and redaction checks for MCP tool calls before they reach downstream servers.
Category
Agent Security
Access Model
Hosted on Monarchic-managed infrastructure
Buyer
Teams letting agents call tools, APIs, terminals, or internal MCPs.
Access
Hosted access is coming soon while accounts, billing, and route access are tested together.
Proof
Check the research and proof boundary before relying on this product for real work.
Account
API keys, hosted MCP config, billing, and account controls will sit behind the product surface.
What You Get
- Hosted MCP route for AgentFirewall
- Scoped API keys for production agents
- Launch timing depends on operational readiness
Core Use Case
Policy-check agent actions before they reach risky tools.
Proof Boundary
Strict MCP input schemas and policy tests cover each public firewall tool.
Outcomes
- Block prompt-injection-shaped tool calls before execution
- Require explicit authority for read, write, execute, and admin actions
- Redact sensitive prompt material and verify provenance claims