# MLSecOpsHub

> MLSecOpsHub is a GitHub organization and community building open-source MLSecOps and AI security tooling, datasets and research for securing ML pipelines, LLM applications, AI agents and autonomous AI architectures. It publishes The MLSecOps Hacker newsletter.

MLSecOpsHub is a GitHub organization and community building open-source MLSecOps and AI security tooling, datasets and research for securing ML pipelines, LLM applications, AI agents and autonomous AI architectures (model supply-chain security, prompt injection, agentic and Model Context Protocol security). The awesome-MLSecOps list has 480 GitHub stars and 29 contributors. The MLSecOps Hacker newsletter on Substack covers the same ground and has published 4 issues.
    

MLSecOps is the practice of applying security engineering to the machine learning lifecycle: training data, model artifacts, pipelines, deployment, and the LLM applications and agents built on top. It extends DevSecOps with AI-specific threats such as model supply-chain attacks through unsafe serialization, prompt injection, and over-privileged agents and MCP tools.

## GitHub projects (public repositories with at least one commit)

- Organization: https://github.com/MLSecOpsHub
- [awesome-MLSecOps](https://github.com/MLSecOpsHub/awesome-MLSecOps): A curated list of MLSecOps tools and resources for securing machine learning and AI systems - adversarial ML defense, LLM security, AI red teaming, model scanning, supply-chain protection, and MLOps pipeline security. Website: https://awesomemlsecops.com/
- [mlsecopshub](https://github.com/MLSecOpsHub/mlsecopshub): MLSecOpsHub website Topics: ai-security, mlsecops, open-source. Website: https://www.mlsecopshub.com/
- [agentic-attack-index](https://github.com/MLSecOpsHub/agentic-attack-index): Open, source-linked dataset of real-world cyberattacks executed or orchestrated by AI agents. Every field cited, every grade honest. Data layer of Rogue Agent Watch. Topics: agentic-ai, ai-agents, ai-security, autonomous-agent-identity, autonomous-agent-security, autonomous-agents, cti, cybersecurity, dataset, llm-security, mitre-atlas, mlsecops, prompt-injection, rogue-agents, rogue-agents-security, sandbox-escape, sandbox-eval, sandbox-evasion, stix, threat-intelligence. Website: https://mlsecopshub.github.io/agentic-attack-index/
- [rogue-agents-watch](https://github.com/MLSecOpsHub/rogue-agents-watch): Rogue Agent Watch: tracking down rogue agents incidents, autonomous AI attacks and AI agents sandbox evasion Topics: agent-identity, agent-security, ai-governance, ai-security-research, ai-security-tool, autonomous-agent-identity, autonomous-agent-security, autonomous-agents, autonomous-agents-security, autonomous-ai-agent-swarms, mitre-attack, owasp, rogue-agents, sandbox-escape, sandbox-eval, sandbox-evasion, security-incidents, threat-intelliegence. Website: https://www.rogueagentswatch.com/
- [provenance-engineering](https://github.com/MLSecOpsHub/provenance-engineering): Open knowledge base for Provenance Engineering: data lineage, identity governance, model provenance and software inventory. Topics: agent-provenance, aibom, cbom, data-lineage, identity, identity-governance, identity-verification, model-provenance, provenance-engineering, sbom, software-inventory.
- [world-models-security](https://github.com/MLSecOpsHub/world-models-security)
- [agentic-access-control](https://github.com/MLSecOpsHub/agentic-access-control): Read-only, local-first dashboard showing what your AI agents (Claude Code, Gemini CLI, MCP servers) are actually allowed to do: permission rules with file provenance, MCP server inventory, sandbox & hook posture, risk findings, OWASP Agentic Top 10 threat scenarios, and config drift. No enforcement, no telemetry (for now). Topics: access-control, agentic-access, ai-agents, ai-agents-security, ai-observability, ai-threat, claude-code-security, gemini-cli, mcp-security, mlsecops, model-context-protocol, owasp, security-posture, threat-modeling.
- [damn-vulnerable-skills](https://github.com/MLSecOpsHub/damn-vulnerable-skills): Vulnerable-by-design security lab for Claude Agent Skills (SKILL.md) and Agent Skills in general Topics: agent-security, agent-skills, ai-security, claude-code, claude-skills, llm-security, mitre-atlas, mlsecops, owasp, skills-registry, vulnerability-research.

## Newsletter

- The MLSecOps Hacker: https://themlsecopshacker.com
- [Research roundup #1: Rogue Agents, Provenance Engineering and Vulnerable Agent Skills](https://themlsecopshacker.com/p/research-roundup-1-rogue-agents-provenance): The AI Security threats and research areas I am researching at the end of this quarter and why
- [Safetensors Won. Model Serialization Attacks Didn’t Stop.](https://themlsecopshacker.com/p/safetensors-won-model-serialization): Winning the format war did not end the supply-chain war.
- [AI Security: Model Serialization Attacks](https://themlsecopshacker.com/p/ai-security-model-serialization-attacks): Navigating the Pitfalls of Model Serialization And ML Supply Chain Vulnerabilities: Best Practices for AI Security
- [What is MLSecOps?](https://themlsecopshacker.com/p/what-is-mlsecops): The machines are coming, but who watches the watchers?
- What is MLSecOps? https://themlsecopshacker.com/p/what-is-mlsecops
- Feed: https://themlsecopshacker.com/feed

## Elsewhere

- X: https://x.com/MLSecOpsHacker
- Tools, resources and glossary: https://awesomemlsecops.com

## Optional

- [Privacy](https://mlsecopshub.com/privacy)
- [Imprint](https://mlsecopshub.com/imprint)
- [Sitemap](https://mlsecopshub.com/sitemap-index.xml)