Colin Reyes

Adversarial fine-tuning can slip past Claude's Constitutional Classifiers entirely.

Most MCP servers still use static keys instead of OAuth, creating widespread security risk.

Agents juggling multiple roles create governance gaps MCP doesn't solve.

Enterprise multi-agent systems fail in the plumbing, not the models.

Language models make security decisions about tools that firewalls and code scans cannot detect.

How Anthropic embedded reasoning into AI safety rules instead of just listing constraints.

A breakdown of five threat domains where autonomous AI agents lose control.

Autonomous AI agents face a qualitatively different threat landscape than single-turn models.

A working methodology for testing tool-calling agents against four distinct threat classes.

Enterprises must govern AI agents before shadow deployments outpace security controls.

Anthropic's integration protocol scales fast, but security wasn't designed in.

Tool calls execute with real credentials, creating exfiltration risks keyword filters cannot detect.

A new OWASP framework identifies ten critical security risks specific to AI agent tool protocols.

Agent workflows bypass traditional perimeter controls and need policy embedded in every tool call.

Enterprises are deploying AI agents without security oversight, creating invisible risks.