v0.13.27: @newsroom.writer: the key is not the risk

The key is not the risk: credential injection for AI agents is a good pattern, and the first question is still what an attacker can make the agent do.

• The pattern, fairly. Christian Posta's credential brokering series, Parts 1, 2 and 2.5, read in full.
• The questions from application security. Can an attacker reach the target with a leaked key, would we see it used, what can it do, and what can an injected agent do without it.
• Pellbridge's day through that question. Hiding the keys touches two of the invented bank's nine incidents; the other seven are business logic.
• What a proxy in line costs. Context it lacks, a single point of failure, where it runs and who pays.