Start with visibility

FuseGov Agent Catalog
the enterprise directory for AI agents

Know what agents are running, who owns them, what tools they can touch, and which ones are safe for production. When you’re ready, enable Active Protection per agent to enforce guardrails at runtime.

See Catalog DemoRegister an Agent

How agents get discovered / registered

Choose the model that fits your org today — and expand later without re-platforming.

Self-register (fastest)
Teams register agents via manifest (agent.yaml) or CLI.
CI/CD registration
Register agents automatically at deploy time (GitHub Actions, GitLab, Azure DevOps).
Platform discovery
Import agents from Kubernetes metadata (labels, namespaces, workload tags).
Roadmap: unmanaged discovery
Network-level discovery for unmanaged traffic in highly regulated environments.
View the Agent Manifest Spec

What you get in Week 1

Ownership & Lifecycle

Every agent has an owner, environment tier (dev/test/prod), risk tier, version history, and an approval trail.

Tool Access Mapping

Define which tools/endpoints an agent can call. FuseGov compiles these settings into policy bundles automatically.

Searchable Discovery

Find the right agent by capability, tags, owner, environment, and risk tier — and reduce duplicate agent builds.

Catalog outcomes (what buyers care about)

Catalog-first metrics make adoption easy, then Active Protection becomes a natural expansion.

Visibility
Shadow agents discovered fast
Surface unmanaged agents and attach ownership in the first week.
Inventory
Complete inventory in weeks
Search by capability, owner, environment tier, and risk classification.
Reuse
Reduce duplicate effort
Help teams find existing agents before building another from scratch.

Fits into enterprise workflows

Catalogs are most valuable when connected to how teams ship. FuseGov plugs into platform workflows — not around them.

CI/CD
Register agents at deploy time; track versions automatically.
Kubernetes
Import identity/context from labels, namespaces, and workload metadata.
Existing catalogs
Link agent records into Backstage / ServiceNow as a first-class service.
Security tooling
Export evidence packs for audit and incident review.

Control plane vs. data plane

The Catalog is the Control Plane. The Enforcement Proxy is the Data Plane. Start with visibility, then turn on control.

Control Plane: Agent Catalog

Manifests, ownership, lifecycle, approvals, and policy bundle generation.

  • • Register agents via UI, API, or Git manifests
  • • Assign owner + environment + risk tier
  • • Tag capabilities and map tools/endpoints
  • • Generate versioned policy bundles

Data Plane: Enforcement Proxy (ASEP)

Deterministic guardrails, semantic checks, degraded mode, evidence packs.

  • • Intercepts tool calls (no agent code changes)
  • • Applies deterministic rules in microseconds
  • • Escalates to semantic analysis only when configured
  • • Produces audit-grade evidence records
Learn about Active Protection

Visibility → Governance → Enforcement

Want the catalog in your environment?

We can onboard your first agents quickly, map tool access, and deliver a working directory that becomes the control plane for enforcement.

Book a Catalog DemoBuyers & Personas
Tip: register agents at deploy time for clean versioning and lifecycle.