Goal
Give automation a typed, bounded-output contract that scales beyond capture-all JSON envelopes.
Background
The JSON/diagnostic model is a strong foundation, but large output and long-running workflows need streaming without unbounded memory while retaining human terminal behavior.
Scope
Define a typed result protocol and schema version; add a bounded NDJSON or equivalent writer; separate machine output from stderr diagnostics; specify ordering, flush, error, and partial-result behavior; retain current JSON compatibility.
Acceptance Criteria
- Large structured output is bounded-memory.
- Records are schema-versioned and documented.
- Human output and diagnostics stay stable.
- Existing JSON consumers continue to work.
- Stream failures yield actionable status.
Validation
Test large output, flush/backpressure, errors, interruption, and legacy JSON; run examples from a clean install and validate the schema.
Non-Goals
Do not remove existing JSON mode or force streaming on commands that do not need it.
Project Fields
- Status: Backlog
- Priority: P1
- Area: CLI
- Initiative: v1.0 Readiness
- Size: L
Ownership
Goal
Give automation a typed, bounded-output contract that scales beyond capture-all JSON envelopes.
Background
The JSON/diagnostic model is a strong foundation, but large output and long-running workflows need streaming without unbounded memory while retaining human terminal behavior.
Scope
Define a typed result protocol and schema version; add a bounded NDJSON or equivalent writer; separate machine output from stderr diagnostics; specify ordering, flush, error, and partial-result behavior; retain current JSON compatibility.
Acceptance Criteria
Validation
Test large output, flush/backpressure, errors, interruption, and legacy JSON; run examples from a clean install and validate the schema.
Non-Goals
Do not remove existing JSON mode or force streaming on commands that do not need it.
Project Fields
Ownership