Prompt Injection Risks in AI Coding Agent Pipelines
Coding agents can't patch prompt injection like databases patched SQL.
Tomás Eguía-Reyes
Contributing Editor
Tomás Eguía-Reyes
Tomás Eguía-Reyes has reported on AI policy and enterprise software governance for more than a decade, with bylines in outlets covering regulatory technology and enterprise IT strategy. His current beat examines the emerging standards, audit frameworks, and principal-hierarchy models being adopted across heavily regulated industries.
3 stories
Per-Session Credential Minting and Revocation in Agent Sandboxes
The architecture that keeps agent credentials from lingering after tasks end.
Tomás Eguía-Reyes
Supply Chain Risks When AI Agents Modify Dependencies
AI dependency recommendations fail at dangerous rates, creating new supply chain attack surfaces.
Tomás Eguía-Reyes