AgentReadyHomeAgent Listing

← Digest AI

Digest AI — agentic threat model

6.2AIVSS 6.2 · Medium

Digest AI is a low-risk, retrieval-focused legal assistant whose primary security concerns are data privacy (leakage of sensitive legal queries) and data integrity (hallucinations or poisoning of legal citations).

OWASP AIVSS score rationale

AIVSS = (CVSS_Base + AARS) × Mitigation_Factor, where AARS = (10 − CVSS_Base) × (Factor_Sum / 10) × ThM
CVSS base 5.3AARS uplift 0.94Factor sum 2.0/10Threat ×1.0Mitigation ×1.0
Autonomy of Action
0.10
Goal-Driven Planning
0.20
Self-Modification
0.00
Dynamic Tool Use
0.20
Persistent Memory
0.20
Contextual Awareness
0.40
Dynamic Identity
0.00
Multi-Agent Interactions
0.00
Non-Determinism
0.50
Opacity & Reflexivity
0.40

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 LLM is not specified. Threats include prompt injection leading to hallucinated legal precedents or citations, and model output misalignment.

L2 · Data Operations✓ mapped

The tool relies heavily on a database of Philippine laws, cases, and precedents. Threats include knowledge-base poisoning (injecting fake precedents) and data exfiltration of proprietary legal documents or user search history.

L3 · Agent Frameworks⚠ not certain from listing

Not certain from the listing — the orchestration framework is unknown. Threats include insecure tool integration between the chat interface and the legal database search engine.

L4 · Deployment & Infrastructure⚠ not certain from listing

Not certain from the listing — hosting details are not provided. Threats include standard web application vulnerabilities, unauthorized access to the vector database, and lack of sandboxing for offline content generation.

L5 · Evaluation & Observability⚠ not certain from listing

Not certain from the listing — no monitoring or guardrail systems are mentioned. Gaps in drift detection or hallucination monitoring could lead to lawyers receiving and using incorrect legal citations.

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

Not certain from the listing — compliance certifications (e.g., SOC2, GDPR/DPA 2012) are not stated. Risks include unauthorized access to sensitive legal queries and lack of audit trails for research activities.

L7 · Agent Ecosystem✓ mapped

The agent operates as a standalone vertical solution with no multi-agent or ecosystem integrations mentioned. Ecosystem threats are minimal, though third-party API dependencies for the LLM could introduce supply chain risks.

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