Spanda

Decision Traceability

Status: Stable — v3 trace emission and proof-field validation tested; envelope crypto signing planned.

Every distributed decision is fully traceable for audit, replay, assurance, diagnosis, and explainability.

Recorded fields

Field	Description
decision_layer	Reflex, local entity, fleet, or control center
entity_id	Originating entity
inputs	Signal values and context at decision time
policy_version	Active policy version
local_context	Offline duration, connectivity state
selected_action	Chosen action
rejected_alternatives	Actions considered but not taken
safety_validation	Safety gate results
trust_validation	Trust gate results
escalation_path	Escalation chain if applicable
outcome	Post-execution result

Security envelope

Every local decision includes:

• Entity identity
• Authority scope
• Policy version
• Decision tree hash
• Timestamp and nonce (replay protection)
• Signature
• Safety and trust validation results
• Audit record ID

Integration

System	Integration
Audit	DistributedDecisionRecord → mutation and mission audit
Replay	Trace files via spanda decision trace
Assurance	Policy version and safety checks in assurance reports
Diagnosis	Decision chain in diagnosis context
Explainability	spanda decision explain plain-language output

CLI

spanda decision trace mission.trace --json
spanda decision explain mission.trace

Enable live emission with SPANDA_DECISION_TRACE=1 during mission execution.

See also decision audit trail.

This site is open source. Improve this page.