Databricks databricks-dabs — agentic threat model
This agent presents a high-risk profile due to its capability to generate and execute Infrastructure-as-Code (IaC) via Databricks Asset Bundles, potentially allowing an attacker to deploy malicious jobs or exfiltrate data if the agent is compromised.
OWASP AIVSS score rationale
| Autonomy of Action | 0.60 | |
| Goal-Driven Planning | 0.50 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.10 | |
| Contextual Awareness | 0.40 | |
| Dynamic Identity | 0.70 | |
| 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.
Not certain from the listing — The underlying foundation model is not specified. Standard LLM risks such as prompt injection could lead to the generation of malicious or malformed bundle configurations.
Not certain from the listing — No specific RAG or vector database is mentioned, though the agent must read and write local bundle configuration files and workspace metadata.
The agent framework orchestrates the translation of user intent into Databricks Asset Bundle (DAB) specs and executes CLI commands. Insecure tool integration or prompt injection could allow an attacker to hijack the CLI flow to execute unauthorized deployment commands.
The agent directly deploys resources (jobs, pipelines) to the Databricks workspace. A compromise at this layer could lead to unauthorized infrastructure provisioning, resource exhaustion, or lateral movement within the Databricks cloud environment.
Not certain from the listing — There is no mention of built-in guardrails, drift detection, or logging mechanisms to monitor the correctness and safety of the generated and deployed configurations.
The agent relies on Databricks CLI credentials and workspace permissions. Strict identity and access management (IAM), least-privilege token scoping, and comprehensive audit logging of the CLI execution are critical to prevent abuse.
Not certain from the listing — While packaged as an 'Agent Skill', there is no explicit detail on multi-agent orchestration or third-party marketplace dependencies that could introduce cascading trust risks.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).