Skip to content

Back to the catalog

The Evolution of JFrog AI Catalog: Your AI Control Plane for Agentic Development
jfrog.comarticle

The Evolution of JFrog AI Catalog: Your AI Control Plane for Agentic Development

JFrog vendor post arguing that agentic development (ADLC) introduces new unvetted supply chains — models, MCP servers, skills/plugins, hooks, and APM packages — that bypass traditional signature-based scanning. It cites real incidents: a malicious MCP server impersonating Postmark's email MCP that BCC'd all agent-sent emails to an attacker, and the 'omnicogg' malicious Claude/agent skill that hid a 22MB encoded payload and evaded all 65 VirusTotal engines for 5,000+ installs, plus references to 2026 supply-chain attacks on LiteLLM, TanStack, and Mistral AI. It then pitches JFrog AI Catalog's expanded MCP/skill/model registries as the fix.

TL;DR Vendor post on JFrog AI Catalog highlights real malicious MCP server and agent-skill supply-chain attacks bypassing signature-based AV scanning.
Open original ↗

Organisations named here

  • JFrog AI Catalog
  • omnicogg

Technologies named here

  • LiteLLM
  • tanstack
  • Mistral AI
  • Claude Code
  • Cursor
  • VS Code
  • OpenCode
  • Codex
  • Artifactory
  • GitHub Copilot
  • VirusTotal
  • Hugging Face
  • NVIDIA NIM
  • APM
  • Model Context Protocol
  • JFrog Artifactory
  • postmark-mcp
  • Agent Guard
  • Agent Packages
  • Agent Plugins Registry
  • Agent Skills Registry
  • JFrog agent plugins
  • JFrog AI Catalog
  • MCP Registry
  • Postmark email MCP (impersonated)
  • Semantic Scan
  • Shadow AI Detection

Techniques named here

  • Artifact signing and verification
  • Encoded payload execution
  • Impersonation of a legitimate package
  • Runtime policy enforcement
  • Semantic scanning
  • Signature-based scanning evasion
  • Software supply chain attack

Concepts named here

  • Agentic Development Lifecycle (ADLC)
  • Agent immunization
  • AI control plane
  • AI supply chain
  • Blast radius
  • Shadow AI
  • Single source of truth

Threat actors named here

  • Malicious Postmark email MCP server impersonation (package that BCC'd every email the agent sent to a threat actor)
  • omnicogg incident (malicious skill hiding a 22MB file with an encoded payload; 5,000+ installs in 19 days; bypassed all 65 VirusTotal engines)

Related in AI Security