Posts tagged "AI agents"

A trajectory grader for detecting unauthorized reads, duplicate effects, fabricated tool results, waste, and unsafe recovery even when an AI agent's final answer looks correct.

A working record-and-replay harness for agent tests: a JSON cassette of decisions and tool results, deterministic replay with zero live calls, and a readable diff when the trajectory changes.

A reference flow for keeping user identity, tenant, audience, and allowed actions attached to every AI tool call without trusting the model to enforce permissions.

When an agent calls a tool, it stops and waits. If the tool is slow, the whole turn is slow, and the user waits with it. The fix is small: put a deadline on every tool call.

Most startups that say they want an AI agent should not start with an agent. Here is how to choose between RAG, workflow automation, and agents — and what to build first.