OVARA.
gateway.
Runtime trust infrastructure for autonomous systems. Ovara intercepts your AI agents at the exact millisecond of action to enforce policy, verify machine identity, and cryptographically audit every decision.
6.12µs
The Missing
Layer.
The Old Model
Cloud IAM, API gateways, and workload identity were built for humans and deterministic microservices: predictable control flow, long-lived credentials, and strictly bounded automation.
The AI Execution Model
Autonomous systems completely break that threat model. Agents make decisions at runtime, use tools adaptively, operate under delegated authority, and drift over long horizons.
Existing infrastructure can authenticate an agent, but it cannot constrain, explain, or revoke them at the exact millisecond they execute a destructive action.
Why OVARA.
Single-binary Go gateway with sub-10μs decision latency
ed25519 machine identity and short-lived capability leases
Trust degradation and sliding-window drift detection
HMAC-SHA256 deterministic execution receipts
11 execution surfaces including shell, Git, GitHub, CI/CD and sandboxed execution
Production hardening: AppArmor, eBPF, Seccomp, Firecracker, Terraform K8s
What breaks
today.
No verifiable identity. Any process can claim to be an agent. Without cryptographic machine identity, impersonation is trivial.
Static keys never expire. A leaked API key works forever: there is no lease, no scope, no blast-radius control.
No record of what ran. After an incident, nobody can prove which action executed and under whose authority.
An interceptor wraps the action across 11 execution surfaces: shell, exec, git.*, github.*, ci.trigger.
action_type, resource, environment, plus optional agent_identity and capability_lease.
Policy evaluation, ed25519 identity verification, lease validation, and trust scoring in a single pass.
The evaluator returns one of three decisions:
Endpoints, fields, and decision values verified against OVARA/runtime/gateway
The 5 Layers
of trust.
Ovara isn't just a proxy. Every intercepted action must successfully pass through five cryptographic and logical primitives before reaching the host system.
Agent
Identity
Enforces ed25519 keypairs. Verifies exactly which machine, container, or sub-agent is requesting the action, eliminating impersonation.
Capability
Lease
Static API keys are dangerous. Ovara issues time-bound, scope-limited leases that automatically expire, minimizing the blast radius of a compromised agent.
Delegation
Chain
Cryptographic proof of inheritance. Tracks exactly which parent agent spawned which sub-agent, maintaining strict provenance across massive autonomous swarms.
Trust
Context
Dynamic scoring engine. Computes sliding-window drift analysis and exponentially degrades trust if environmental anomalies or policy violations are detected.
Execution
Receipt
A deterministic, HMAC-SHA256 hashed log proving exactly what ran, when it ran, and under whose cryptographically verified authority. Perfect for compliance.
Runtime
Interceptor.
// Evaluation Flow //
Sub-10μs decisions.
Fast enough for inline interception in agent workflows. Every decision runs identity verification, lease validation, policy evaluation, and trust scoring in microseconds.
Under
Construction.
Ovara is in active development toward a V1.0.0 release. The gateway compiles to a single, static Go binary. Follow the repository for release updates.
$ curl -X POST http://localhost:8080/v1/runtime/check \-d '{ "action_type": "shell","resource": "shell:git push origin main","environment": "dev" }'# -> { "decision": "escalate",# "requires_approval": true,# "trust_score": 0.5, "receipt_stub": {...} }
$ git clone https://github.com/SidianLabs/OVARA.git$ cd OVARA/runtime/gateway$ go build -o ovara-gateway ./cmd/server$ ./ovara-gateway
Block or approve AI-driven shell commands and infrastructure changes
Enforce least-privilege capability leases across agent fleets
Detect anomalous behavior with sliding-window drift analysis
Generate cryptographic receipts for compliance and forensics