Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

Plugin-authoring guidance caps skill length at 500 lines but sets no maximum line length

Aperta
  • Qualcosa non va?
#113 1 commento 2 reazioni 0 assegnatari Vedi su GitHub

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
5/5
Tempo stimato
Più di una settimana
Idoneità per principianti
35/100
Tipo di issue
Documentazione
Chiarezza
Da chiarire
Stato di attività
Tranquilla
Ambito
documentation

Direzione di ricerca

Start by comparing the 500-line guidance in progressive-disclosure.md and cowork-specific-skill-instructions.md with agent-building-guidelines/agent-description-length.md. Decide whether to use line, character, or token limits, which files they cover, and whether they are enforced or recommended. Done means the guidance has a consistent, explicit policy for skills, agents, and references.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

Summary

The plugin-authoring guidance caps SKILL.md length at 500 lines but sets no maximum line length. Agent guidance is looser still: it has no body line-count cap or line-length rule at all, only a frontmatter description character budget.

Where the cap lives

  • han-plugin-builder/skills/guidance/references/skill-building-guidance/progressive-disclosure.md (line 36) — "keep the SKILL.md body under 500 lines... Treat 500 lines as the ceiling, not the target." No line-length rule.
  • han-plugin-builder/skills/guidance/references/skill-building-guidance/cowork-specific-skill-instructions.md (lines 167, 249) — repeats the 500-line cap. No line-length rule.
  • han-plugin-builder/skills/guidance/references/agent-building-guidelines/ — the only length guidance is agent-description-length.md (description char budget: 1024 target, ~1500 hard signal). No body line-count or line-length rule for agents.

Why it matters

The 500-line cap exists to bound how much content a skill loads into the model's context window and to keep bloaty skills from distracting agents. But a line is unbounded in length, so a line-count cap alone is a leaky proxy for content volume: 500 very long lines can blow past the intended budget. Without a line-length bound (or a more direct measure), the 500-line ceiling doesn't reliably do the job it exists for.

Open question

Not just "what max line length" — the real question is which lever is right:

  • a per-line character cap (e.g. 80 / 100 / 120, or semantic line breaks / one sentence per line), or
  • a total character or token budget that measures content volume directly, or
  • keeping the line-count cap and adding a line-length bound so it stops being a leaky proxy.

Scope to decide alongside it: does it apply to agent bodies and references/ files too, not just SKILL.md? Hard cap or recommendation? A judgment call, or something a formatter/linter enforces?

Lingua principale
Shell
Stelle
282
Fork
24
Merge medio
6h 6m
PR unite (30g)
1

Preparare l'ambiente

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di testdouble/han

Tutte le issue di testdouble/han

Issue simili

Altre issue su Shell/Bash

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.