ketraterm-input-encoder

name	ketraterm-input-encoder
description	Input encoder guidance for this Kotlin terminal repository. Use when adding or changing ketraterm-input keyboard, paste, focus, mouse, host-output, or input mode-bit behavior.

Terminal Input Encoder

Use this skill for work in :ketraterm-input or shared API needed by input.

Workflow

1. Read root AGENTS.md, then ketraterm-input/AGENTS.md.
2. Follow ketraterm-input/docs/ketraterm-input-implementation-plan.md.
3. Keep TerminalHostOutput in :ketraterm-protocol.
4. Read packed mode bits once per event through core's input-readable API.
5. Add tests before or alongside each behavior slice.
6. Keep mouse deferred until keyboard, paste, and focus are implemented and passing.

Boundaries

• Input encodes host-bound bytes; it does not parse terminal output.
• Input may depend on protocol vocabulary and core API mode reads.
• Input must not depend on parser, host, grid storage, cursor internals, renderer state, or UI toolkit types.
• Do not decode mode bit positions inside input. Add missing helper methods to the core API first.

Hot Path Rules

• Use primitive bit masks for modifiers.
• Use reusable scratch buffers for dynamic CSI/SS3 sequences.
• Avoid StringBuilder, sliceArray, regex, or allocation-heavy helpers in encoder hot paths.
• writeBytes(bytes, offset, length) consumers must synchronously consume or copy the byte range because encoders may immediately reuse buffers.

Testing

Assert exact bytes for every encoded event. Include validation failures, modifier CSI parameters, UTF-8 printable codepoints, Ctrl/Alt combinations, application cursor/keypad modes, bracketed paste wrappers, focus reporting, and one real core mode-bit host case.