Cykel AI — agentic threat model
Cykel AI presents a high agentic risk profile due to its fully autonomous, 24/7 digital workers operating without human supervision across sensitive business domains like HR, sales, and research, with deep integration into existing corporate systems.
OWASP AIVSS score rationale
| Autonomy of Action | 0.90 | |
| Goal-Driven Planning | 0.80 | |
| Self-Modification | 0.20 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.60 | |
| Contextual Awareness | 0.70 | |
| Dynamic Identity | 0.50 | |
| Multi-Agent Interactions | 0.60 | |
| Non-Determinism | 0.70 | |
| Opacity & Reflexivity | 0.60 |
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.
Not certain from the listing — No details are provided regarding the underlying foundation models used by Lucy, Eve, or Samson, leaving them potentially vulnerable to standard LLM threats like prompt injection or adversarial manipulation.
Not certain from the listing — The mechanisms for data ingestion, vector storage, and RAG are unspecified, raising potential risks of data exfiltration or knowledge-base poisoning given their access to HR and sales data.
High risk of tool misuse and insecure integration. The agents are designed to perform complex business tasks autonomously and integrate with existing systems, meaning compromised orchestration could lead to unauthorized API execution in external HR or CRM platforms.
Not certain from the listing — The deployment architecture, sandboxing of agent actions, and secret management for system integrations are not detailed in the public directory.
Not certain from the listing — There is no mention of real-time monitoring, guardrails, or logging frameworks to detect anomalous agent behavior or drift in their 24/7 autonomous operations.
Not certain from the listing — Compliance alignments (such as SOC2, GDPR for HR data, or ISO) and identity/access management policies for the digital workers are not specified.
The platform hosts multiple specialized digital workers (Lucy, Eve, Samson). While they target different business functions, their co-existence on a single platform introduces potential risks of cross-agent trust abuse or cascading failures if one agent is compromised.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).