Journal topic
Design.
Why Mandare is built the way it is: design choices, their trade-offs, and what they do not cover.
-
LLM observability vs audit trail: what a trace can't prove
Langfuse and LangSmith traces are built for debugging. What their docs say about deleting and expiring traces, and what an audit trail needs instead.
-
AI agent audit trail: why the agent can't write it
An agent's own account of its actions is testimony, not evidence. What the research shows, and where an audit trail has to be written instead.
-
Fail closed or fail open: what an AI gateway should do
What a gateway does with an agent's request when its budget store, ledger or approver is down: Envoy, Stripe and LiteLLM defaults, and which way to fail.
-
Proofs, not data
Why the accountability layer for AI agents must not see your data — and how content-free fingerprints still make the record checkable by anyone.