Agent Maturity Compass API - v1.2.0
    Preparing search index...

    Interface MemoryEvent

    Memory Integrity Scoring — Does the agent maintain consistent context?

    Based on Reddit community insight: "Agent demos look great. Then they fail quietly without a memory layer." This module scores memory health across four dimensions.

    Dimensions:

    • Consistency: Does the agent contradict itself across sessions?
    • Decay: Does context quality degrade over time?
    • Poisoning resistance: Can external input corrupt memory?
    • Recovery: Can the agent rebuild context after loss?
    interface MemoryEvent {
        confidence: number;
        key: string;
        sessionId: string;
        source: "agent" | "system" | "user" | "external";
        timestamp: number;
        type: MemoryEventType;
        value?: string;
    }
    Index
    confidence: number
    key: string
    sessionId: string
    source: "agent" | "system" | "user" | "external"
    timestamp: number
    value?: string