Queryable Infrastructure

Agentic operations require queryable infrastructure, and a benchmark on aws-bench showed that toolchains reading stored copies of the estate scored 0 out of 3 when asked which security groups were attached to nothing. The live cloud can be queried like a database, and chant tooling queries TypeScript as data through the syntax tree, reducing fragmentation between live systems and stored copies.

Queryable Infrastructure Agentic ops means agents need to query your infra on demand. If the live details of your estate are fragmented this can lead to extra lookups costing time and tokens. I measured this on aws-bench /posts/aws-bench-scenario-1-wrap/ : asked which security groups were attached to nothing, every toolchain that reads a stored copy of the estate scored 0 out of 3. Many things can be queried like a database without being one. The live cloud is one of them. Typed source is another. chant tooling reads TypeScript as data https://intentius.io/chant/concepts/typescript-as-data/ through the syntax tree, so an agent queries your infra code instead of grepping it. The fragmentation of truth between the live system and a stored copy is far more serious than IaC spread across repos. When many repos deploy changes they converge in the live system, which is already authoritative and already queryable. The repo layout never enters into it. Agents benefit from queryable infrastructure, and the spec covers most of it /posts/the-long-tail-is-a-tail/ .