ML//agent//agent harness//hidden ABI

The hidden ABI is the looser, probabilistic contract between an LLM and its harness, the agent's equivalent of the binary ABI that tells compiled components how to exchange arguments and results.

The hidden ABI is the looser, probabilistic contract between an LLM and its harness, the agent's equivalent of the binary ABI that tells compiled components how to exchange arguments and results.

This is an editorial analogy, not a claim that agent systems lack prior interface concepts. Agent-Computer Interface already names the broader interaction surface. Hidden ABI isolates the behavioral fact that a representation can remain software-compatible while changing how a model selects and interprets actions.

Its effective contract includes tool names, descriptions, schemas (the tool interface, its most visible part), examples, output formatting, truncation, and error semantics. Changing only a description can change which action the model selects even when the implementation remains byte-for-byte identical.

P(a∣c,T)P(a\mid c,T)P(a∣c,T)

Here the available tools TTT are not passive documentation. They modify the policy over actions. This makes the interface partly executable: natural language and data structures program behavior without deterministically specifying it.

Traditional compatibility asks whether a call still runs. Agent compatibility must also ask whether the model still understands when to make the call.

Action representationNames, descriptions and schemas alter tool selection.

Observation representationOutput shape, identifiers and truncation alter the next decision state.

Recovery representationErrors teach the model which retry or fallback remains available.

The adversarial consequence is measurable: ToolTweak moved selection from roughly 20% to as high as 81% by manipulating tool names and descriptions.