AgentReadyHomeAgent ListingPricing

← Harpa AI

Harpa AI — agentic threat model

9.1AIVSS 9.1 · Critical

Harpa AI presents a high-risk profile primarily due to its deployment as a browser extension with deep DOM access, session context awareness, and the ability to execute custom automation commands, making it a prime target for data exfiltration and session hijacking.

OWASP AIVSS score rationale

AIVSS = (CVSS_Base + AARS) × Mitigation_Factor, where AARS = (10 − CVSS_Base) × (Factor_Sum / 10) × ThM
CVSS base 8.3AARS uplift 0.82Factor sum 4.6/10Threat ×1.05Mitigation ×1.0
Autonomy of Action
0.60
Goal-Driven Planning
0.40
Self-Modification
0.10
Dynamic Tool Use
0.70
Persistent Memory
0.30
Contextual Awareness
0.80
Dynamic Identity
0.50
Multi-Agent Interactions
0.10
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✓ mapped

Integrates with external foundation models (GPT, Claude). Threats include prompt injection via web page content (indirect prompt injection) leading to model reprogramming or mis-aligned outputs during summarization.

L2 · Data Operations✓ mapped

Processes active browser page data and performs automated data extraction. Threats include data exfiltration of sensitive session data, and knowledge-base poisoning if malicious web content is ingested during monitoring.

L3 · Agent Frameworks✓ mapped

Orchestrates customizable automation commands and web monitoring. Threats include tool misuse, where malicious page elements trigger unintended automation commands or exploit insecure tool integrations.

L4 · Deployment & Infrastructure✓ mapped

Deployed as a browser extension. Threats include extension-level privilege escalation, unauthorized DOM access, and local storage compromise of API keys or session tokens.

L5 · Evaluation & Observability⚠ not certain from listing

Not certain from the listing — no details are provided regarding logging, guardrails, or drift detection for the browser-based automation and LLM interactions.

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

Not certain from the listing — no compliance certifications (e.g., SOC2), identity governance, or specific data privacy controls are detailed in the public directory.

L7 · Agent Ecosystem⚠ not certain from listing

Not certain from the listing — there is no mention of multi-agent coordination, agent-to-agent trust boundaries, or marketplace integrations.

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

These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.