AgentReadyHomeAgent Listing

← Clippy AI

Clippy AI — agentic threat model

6.9AIVSS 6.9 · Medium

Clippy AI presents a moderate security risk primarily driven by its Slack integration and RAG-based knowledge base, which are susceptible to prompt injection and data exfiltration, though mitigated by human-in-the-loop handoffs for complex queries.

OWASP AIVSS score rationale

AIVSS = (CVSS_Base + AARS) × Mitigation_Factor, where AARS = (10 − CVSS_Base) × (Factor_Sum / 10) × ThM
CVSS base 6.8AARS uplift 0.9Factor sum 2.8/10Threat ×1.0Mitigation ×0.9
Autonomy of Action
0.50
Goal-Driven Planning
0.20
Self-Modification
0.10
Dynamic Tool Use
0.30
Persistent Memory
0.20
Contextual Awareness
0.40
Dynamic Identity
0.10
Multi-Agent Interactions
0.10
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 foundation model is not specified. Standard LLM risks apply, including prompt injection that could bypass canned responses or cause the bot to output misaligned or brand-damaging content.

L2 · Data Operations✓ mapped

The agent relies heavily on a user-provided knowledge base (FAQs and company info). This introduces risks of knowledge-base poisoning if unauthorized users can modify the FAQs, or data exfiltration if sensitive internal documents are uploaded and leaked via prompt injection.

L3 · Agent Frameworks✓ mapped

The orchestration framework manages conversation flow, canned responses, and human handoff. Vulnerabilities include prompt injection manipulating the handoff logic (e.g., preventing handoff to keep a user trapped with the bot, or forcing premature handoffs to flood human agents).

L4 · Deployment & Infrastructure⚠ not certain from listing

Not certain from the listing — As a closed-source SaaS platform, the hosting, sandboxing, and network isolation details are unknown. Risks include tenant isolation failure or insecure API endpoints connecting the chatbot to the Slack workspace.

L5 · Evaluation & Observability⚠ not certain from listing

Not certain from the listing — While 'AI Training' is mentioned, there is no explicit detail on real-time guardrails, logging, or drift detection. This creates a blind spot where malicious interactions or prompt injection attempts may go undetected.

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

Not certain from the listing — No compliance certifications (e.g., SOC2, GDPR) or role-based access controls (RBAC) for the admin panel are detailed. A lack of strong admin authentication could allow attackers to alter the bot's training data.

L7 · Agent Ecosystem✓ mapped

The agent integrates directly into the Slack ecosystem. A compromise of the Slack workspace or a rogue user within the workspace could abuse the bot to extract proprietary knowledge base data or conduct internal phishing campaigns.

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