Back to Architecture
Protection Layer
Safety Architecture
Defense-in-depth protection with six layers ensuring agents operate safely within defined boundaries.
Six Protection Layers
Circuit Breaker Protocol
Last-line defense against irreversible mistakes. Blocks all IRREVERSIBLE actions until approved.
Reversibility Assessment
Evaluates actions as IRREVERSIBLE (blocked), PARTIALLY_REVERSIBLE (approval needed), or FULLY_REVERSIBLE (proceed).
Impact Assessment
Data loss risk, security impact, system availability, recovery plan validation.
Bounded Autonomy
Tier-based capability restrictions. Tier 1 (full RIGOR), Tier 2 (R+I+G), Tier 3 (Research+Inspect only).
Human Approval Workflows
ACMF mode-based and risk-level-based approval requirements with 24-hour timeouts.
Audit Trail & Compliance
Universal audit logging for EU AI Act compliance. Every action logged with decision reasoning.
Reversibility Assessment
IRREVERSIBLE→ Always blocked
PARTIALLY_REVERSIBLE→ Requires approval
FULLY_REVERSIBLE→ Proceed (subject to other controls)
Circuit Breaker Protocol
// Circuit Breaker States
type CircuitState string
const (
CircuitClosed CircuitState = "closed" // Normal operation
CircuitOpen CircuitState = "open" // Blocking all requests
CircuitHalfOpen CircuitState = "half_open" // Testing recovery
)
// Circuit Breaker Configuration
type CircuitBreaker struct {
Name string
State CircuitState
FailureCount int
SuccessCount int
FailureThreshold int // 5 failures to open
SuccessThreshold int // 3 successes to close
Timeout time.Duration // Time before testing recovery
LastFailure time.Time
OnStateChange func(from, to CircuitState)
}
// Cascade Protection
// If service A fails → Circuit opens → Service B protected
// After timeout → Half-open → Test recovery
// Success → Close circuit → Normal operationBounded Autonomy Tiers
| Tier | RIGOR Phases | Capabilities |
|---|---|---|
| Tier 1 (Full) | R + I + G + O + R | Self-Healing, Self-Deploying, Self-Protecting |
| Tier 2 (Basic) | R + I + G | Self-Scaling, Self-Optimizing, Self-Diagnosing + 9 more |
| Tier 3 (Safety) | R + I | Self-Monitoring, Code Analysis, Compliance + 9 more |