AI security guide · MITRE ATLAS
Model, dataset and agent-tool supply chain
What are AI supply-chain threats? AI systems depend on external models, datasets, libraries and agent tools. Any of these can be poisoned, backdoored or replaced after adoption ("rug pull").
Why it matters
Agent tool ecosystems and model hubs make it easy to add capabilities with a single install — and equally easy to add an adversary’s code with broad access.
How it works
- Publish poisoned models, datasets or tools to public repositories.
- Compromise a dependency used in AI pipelines.
- Change a previously benign tool’s behaviour after adoption.
Curated agent example
Fictional example: a third-party agent tool
- A team installs a community tool that "summarizes cloud costs".
- A later version also reads environment variables and sends them externally.
- Agent credentials are exposed.
- Mitigation: pin and review tool versions; run tools with isolated, minimal credentials and restricted egress.
Fictional and simplified for education.
On AWS: architecture notes
- Host approved models and artifacts in controlled registries (for example private ECR / S3 with signing).
- Run third-party tools in isolated execution roles and VPCs with restricted egress.
MITRE ATLAS techniques
- AML.T0010 Supply Chain Compromise
Initial Access · Maturity: Realized
Adversaries may manipulate products, components, or their development and delivery mechanisms before they reach downstream consumers to compromise data or systems. Manipulation may involve source code, dependencies, development tools, build environments, software updates, distribution channels, or hardware. In AI environments, targets may include Hardware, Software, Data Repositories, Model Repositories, Container Registries, and AI Agent Tools. Compromised components may introduce malicious executable behavior, alter model behavior, or provide unauthorized access when incorporated into downstream systems.
- AML.T0010.001 Supply Chain Compromise: Software
Initial Access · Maturity: Realized
Adversaries may target software packages that are commonly used in AI-enabled systems or are part of the AI DevOps lifecycle. This can include deep learning frameworks used to build AI models (e.g. PyTorch, TensorFlow, Jax), generative AI integration frameworks (e.g. LangChain, LangFlow), inference engines, and AI DevOps tools. They may also target the dependency chains of any of these software packages. Additionally, adversaries may target specific components used by AI software such as configuration files or example usage of AI packages, which may be distributed in Jupyter notebooks.
- AML.T0010.003 Supply Chain Compromise: Model Repository
Initial Access · Maturity: Realized
Adversaries may abuse model repositories or other model distribution services to introduce compromised models into downstream AI systems. They may replace published artifacts, manipulate model versions or references that consumers resolve, or publish malicious models under identities or locations that consumers trust.
- AML.T0110 AI Agent Tool Poisoning
Persistence · Maturity: Realized
Adversaries may poison tools used by AI agents by introducing or modifying malicious content or behavior in a tool's model-visible definition, executable implementation, or runtime responses. Tools may include built-in integrations, locally installed packages, remotely hosted services, Model Context Protocol (MCP) servers and tools, and agent skills or similar capability packages.
- AML.T0109 AI Supply Chain Rug Pull
Defense Evasion · Maturity: Realized
Adversaries may publish legitimate AI components or software, gain user adoption, then push an update with a malicious variant, leading to Supply Chain Compromise. More scrutiny is often placed on a supply chain dependency when it is first being considered for inclusion in an AI system. Performing a rug pull may allow adversaries to bypass these defenses and be more likely to achieve Initial Access. Adversaries may Publish Poisoned AI Artifacts, then attempt to gain user trust and increase adoption before performing the rug pull (See AI Supply Chain Reputation Inflation).
- AML.T0115 Publish Poisoned AI Artifacts
Resource Development · Maturity: Realized
Adversaries may create or modify AI artifacts and publish them through public or shared distribution channels to facilitate compromise of downstream AI systems. Poisoned AI artifacts may include datasets, models, and AI agent tools containing malicious content, behaviors, code, or configurations. Adversaries may publish novel artifacts or malicious variants of legitimate artifacts through dataset or model repositories, package registries, source code repositories, tool hubs, or remotely hosted services. Victims may subsequently acquire and integrate these artifacts through Supply Chain Compromise.
ATLAS mitigations
- AML.M0014 Verify AI Artifacts
Verify the integrity and authenticity of AI artifacts before importing, loading, or deploying them. Artifacts may include model weights and checkpoints, datasets, tokenizers, model configurations, prompt templates, and agent configuration files. Verify the complete set of files used by the AI system, including supporting files that can alter model or agent behavior.
- AML.M0023 Maintain an AI Bill of Materials
Create and maintain an AI Bill of Materials (AI BOM) that identifies the components and dependencies used to build, evaluate, and operate an AI-enabled system. Include models and checkpoints, datasets, software packages, inference runtimes, container images, and agent configurations including system instructions and connected tools and data sources. For systems that use external AI services or agent tools, record the relevant providers, services, integrations, and available version identifiers.
- AML.M0013 Code Signing
Enforce binary and application integrity with digital signature verification to prevent untrusted code from executing. Adversaries can embed malicious code in AI software or models. Developers should also cryptographically sign SBOM and AIBOM components that track model or data provenance. Enforcement of code signing can prevent the compromise of the AI supply chain and prevent execution of malicious code.
- AML.M0016 Vulnerability Scanning
Vulnerability scanning is used to find potentially exploitable software vulnerabilities to remediate them.
Related Enterprise ATT&CK techniques
- T1195 Supply Chain CompromiseATLAS-published reference: ATLAS publishes T1195 as the ATT&CK reference for this technique.
- T1195.002 Supply Chain Compromise: Compromise Software Supply ChainRelated enterprise behavior in the same attack chain (Securitain analysis). Not an ID equivalence.
Controls to prioritize
Restricted agent egress and connectors
Limit the domains and services an agent runtime can reach, avoid generic "fetch any URL" tools, and approve third-party connectors before use.
MITRE ATLAS mitigation: Segmentation of AI Agent ComponentsDedicated least-privilege identity per agent and tool
Give each agent tool its own narrowly scoped role (or a session policy) instead of a shared broad role, so a manipulated agent can only do what that tool legitimately needs.
Security best practices in IAMSecret scanning in code and pipelines
Scan repositories, build logs, container images and infrastructure templates for embedded credentials before they ship, and treat any detected AWS key as compromised.
OWASP Secrets Management Cheat Sheet
Common misconceptions
- “Popular models and tools are safe.”
- Popularity is not provenance; verify sources and pin versions.
Checklist
- Maintain an AI bill of materials
- Pin and review model/tool versions
- Isolate third-party tools
Related attack scenarios
Compromised deployment pipeline role
AI agent with excessive AWS permissions
More AI security guides
References
MITRE ATT&CK® and ATT&CK® are registered trademarks of The MITRE Corporation. ATT&CK data © 2026 The MITRE Corporation. This work is reproduced and distributed with the permission of The MITRE Corporation (ATT&CK v20.1). MITRE ATLAS™ data (v2026.10) © MITRE, used under the Apache License 2.0. Securitain is not affiliated with or endorsed by MITRE. Explanations, scenarios and control mappings on this page are Securitain’s own analysis. Sources and licensing