author-intent-yaml

To convert interview notes into an intent.yaml file, provide a structured interview digest containing the intent statement, constraints, and failure conditions. The system processes these inputs to generate a schema-conforming YAML artifact with deterministic positional IDs and strict formatting, excluding implementation details.

The ICE clean triple for intent drafting workflows consists of the intent statement, constraints, and failure conditions. It isolates the core intent from execution mechanisms, ensuring scenarios and recovery steps are excluded and generated later in downstream expectation.yaml artifacts.

Deterministic constraint and failure condition IDs are generated by assigning positional identifiers, specifically C1..Cn for constraints and F1..Fn for failure conditions. This enforces strict schema formatting and ensures stable references across the generated intent.yaml artifact.

Yes, intent.yaml schema generation blocks implementation details. It enforces guardrails that reject text referencing execution mechanisms like plays, agents, skills, tools, or internal framework paths, ensuring the intent remains completely execution-neutral.

Authoring intent.yaml artifacts requires a structured interview digest input containing the name, description, intent_statement, constraints, failure_conditions, and output_base. This structured input is processed to output the YAML in the required top-level order.

Scenarios and recovery steps should be excluded from intent drafting when producing the intent.yaml artifact. These elements are generated later into a separate expectation.yaml file, keeping the initial intent section focused solely on statements, constraints, and failure conditions.