🧰 CrewAI

The items that name it 13

• VIPER-MCP: Detecting Vulnerabilities in MCP Servers
  paperas tools2026-09-17

  Agent development framework that has adopted MCP as a first-class integration layer.

• Top Agentic AI security resources — September 2026
  articleas affected products2026-09-04

  One of the frameworks affected by the eleven reported vulnerabilities

• Awesome Cybersecurity Agentic AI
  repoas tools2026-08-31

  Open-source framework for orchestrating teams of AI agents including security contexts

• AI Development and Agentic Security Labs
  courseas tools2026-08-21

  Referenced for multi-agent agentic pattern labs.

• AI Security Briefing
  writeupas affected products2026-07-18

  Agentic framework cited as affected by a separate RCE chain finding referenced in this article

• GoogleCloudPlatform/k8s-aibom
  repoas affected products2026-07-18
• awesome-ai-security-tools
  repoas affected products2026-07-11

  agentic workflow framework covered by inkog and agentic-radar

• When Prompts Become Shells: RCE Vulnerabilities in AI Agent Frameworks (Semantic Kernel)
  writeupas tools2026-05-07

  AI agent framework mentioned as similarly foundational to the AI agent ecosystem (subject of future research in the series)

• ToFu: Exploiting and Fuzzing the LLM–Orchestrator–Tool Trust Boundary
  paperas classification.technologies
• santosomar/AI-agents-for-cybersecurity
  repoas toolsIt is written here as CrewAI: vulnerability management crew with nmap and CVE tools.
• AI Agents for Cybersecurity: A Guided Tour of a Repository with Tons of Resources and Examples
  articleas toolsIt is written here as CrewAI: four-agent YAML-defined vulnerability-management crew; nmap tool with allowlist target validation, NVD API CVE tool with timeout.
• AI Agents for Cybersecurity: A Guided Tour of a Repository with Tons of Resources and Examples
  articleas classification.technologies
• Delegation Without Trust: An Empirical Gap Analysis of Identity, Authorization, and Runtime Governance in Multi-Agent LLM Systems
  paperas affected products

  multi-agent LLM framework; provides no built-in confinement per the paper's evaluation

In every workspace you can read 23