abcd's help fits the screen it is read on

Press Release

A person reads abcd --help, or abcd --help --agent, in a laptop terminal or a narrow window and every line fits the screen. Each verb's name stands on its own, its one-line purpose wraps to the terminal's width on a word boundary, and what it writes and what it refuses sit on the lines below, so nothing runs off the edge or breaks mid-word when the terminal wraps it. At 80 columns the whole menu reads top to bottom without scrolling sideways, and a wider terminal uses the width it has.

Why This Matters

Typed links: refines itd-146 (the labelled command groups this help renders in) and itd-2610090831227812 (the person's menu and the --help --agent expansion whose lines this re-lays).

The help is the first thing a person reads, and today it cannot be read on an ordinary screen. Measured on 2026-10-30 with COLUMNS=80: abcd --help runs to 64 lines whose longest is 172 characters, and abcd --help --agent to 94 lines whose longest is 218. Every verb is one line holding its name, its purpose, what it writes, what it refuses and, in the agents block, the page to read next ((read commands/<verb>.md)), so the terminal wraps each line at its edge, mid-word and under the name column, and the part that says what a verb writes or refuses lands in an unaligned tail. The width the help is laid out to is the terminal's, never a fixed one, so a wide terminal is not squeezed either.

Whether --help --agent should write JSON by default was researched on 2026-10-30 and answered no: the agents block stays text, read by a model the way the host reads a skill listing, with JSON behind --json. So this draft covers the text layout of both listings, which an agent reads too, and the machine-readable listing is iss-2610301320457709.

Mechanism

Prompted (the claim-recording gradient): why the authors expect this to work, as a falsifiable "we expect X because Y" — not the outcome restated. Replace this line with the claim, or with the exact token None stated. alone on its line to record the claim as considered and declined.

Scope Conditions

Required (the claim-recording gradient): the population, platform, scale, or assumptions this claim holds under, one per top-level bullet — abcd intent plan stamps each with a persistent identity. Replace this line with those bullets, or with the exact token None stated. alone on its line.

Acceptance Criteria

Required (the itd-1 discipline): add at least one Given-When-Then bullet describing the verifiable bar for "shipped" before this draft can be planned.

Open Questions

None recorded yet.

Audit Notes

Empty. Populated by intent-auditor when intent moves to shipped/.