MAGIC Agent Skills is now open source! Star on GitHub
MAGIC Agent SkillsMAGIC Agent Skills
Commands

/linguistic:explore

Investigate a corpus or language interactively, without committing to a pipeline plan.

Syntax

/linguistic:explore [language or corpus]

Description

Enter exploration mode — read-only, curiosity-driven inspection of a language or corpus before deciding what to do. The orchestrator will read available data and surface:

  • Language identification confidence (via linguistic-scope)
  • Script and encoding observations (via linguistic-scripts)
  • Corpus size, register mix, dedup stats (via linguistic-corpus)
  • Tokenizer fertility against a baseline (via linguistic-tokenize)

You stay in exploration mode until you opt into a plan via /linguistic:propose or /linguistic:lifecycle.

Options

ArgumentDescription
[language]Target language name or ISO code
[corpus path]Path to a corpus file or directory to explore

Example

/linguistic:explore Yoruba
> Exploring Yoruba (yor)...
ISO: yor | Glottolog: yoru1245 | Script: Latin + tone diacritics
Resource class: Joshi 2 (Hopefuls)
Fertility on tiktoken-cl100k_base: 3.4× — vocab extension recommended
Corpus available: Bible-NLP (CC-BY), OPUS, CulturaX subset
Register: liturgical 45%, web 35%, news 20% — Bible% high; flag
  • /linguistic:propose — generate a plan based on exploration findings
  • /linguistic:lifecycle — enter full pipeline
Was this page helpful?
Edit on GitHub

Last updated on

On this page