Will attackers successfully compromise AI credentials or MCP systems via scanning campaign by Sept 2026?

SANS ISC (Jul 13) reports active reconnaissance targeting MCP servers and AI assistant credentials. Attackers are scanning for exposed endpoints to harvest credentials used by AI tools and inference systems. This signals escalating threat-actor focus on AI supply-chain and credential theft.

Resolution Criteria

Publicly disclosed breach or compromise of credentials (API keys, authentication tokens) or MCP system access linked to the scanning campaign. Evidence from vendor advisories, breach disclosures, or threat intelligence reports naming the campaign.