AgentReadyHomeAgent Listing

← Tanka

Tanka — agentic threat model

8.7AIVSS 8.7 · High

Tanka presents a high data-security risk profile due to its core feature of long-term enterprise memory and tool integration, which increases the impact of memory poisoning and data exfiltration if compromised.

OWASP AIVSS score rationale

AIVSS = (CVSS_Base + AARS) × Mitigation_Factor, where AARS = (10 − CVSS_Base) × (Factor_Sum / 10) × ThM
CVSS base 7.5AARS uplift 1.18Factor sum 4.7/10Threat ×1.0Mitigation ×1.0
Autonomy of Action
0.40
Goal-Driven Planning
0.30
Self-Modification
0.20
Dynamic Tool Use
0.50
Persistent Memory
0.90
Contextual Awareness
0.80
Dynamic Identity
0.20
Multi-Agent Interactions
0.30
Non-Determinism
0.60
Opacity & Reflexivity
0.50

Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.

MAESTRO 7-layer threat model

Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.

L1 · Foundation Models⚠ not certain from listing

Not certain from the listing — The underlying foundation models are not specified, leaving the system vulnerable to standard LLM threats like adversarial prompt injection and model-specific alignment bypasses.

L2 · Data Operations✓ mapped

Tanka heavily relies on long-term memory and learning from past conversations, making it highly susceptible to data/knowledge-base poisoning and unauthorized data exfiltration of sensitive enterprise context.

L3 · Agent Frameworks✓ mapped

The agent framework manages long-term memory and tool integrations. Insecure tool integration and memory poisoning are primary threats, as malicious chat inputs could manipulate the agent's stored context and tool execution.

L4 · Deployment & Infrastructure⚠ not certain from listing

Not certain from the listing — No deployment details, hosting environments, or sandboxing mechanisms are described, presenting potential risks of container compromise or unauthorized lateral movement.

L5 · Evaluation & Observability⚠ not certain from listing

Not certain from the listing — There is no mention of evaluation frameworks, real-time monitoring, or guardrails to detect drift, anomalous memory updates, or malicious inputs.

L6 · Security & Compliance (cross-cutting)⚠ not certain from listing

Not certain from the listing — Compliance certifications (e.g., SOC2, GDPR) and enterprise access controls are not detailed, which is a critical gap for a tool acting as an 'enterprise brain'.

L7 · Agent Ecosystem⚠ not certain from listing

Not certain from the listing — While designed for team chat environments, it is unclear if Tanka interacts directly with other autonomous agents or if it operates solely as a single-agent assistant.

MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).