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

Commands Reference

The Linguistic Agent Skills suite provides 10 slash commands. Every command has a natural-language equivalent — the linguistic-orchestrator routes both identically.

Command Overview

CommandPurposeNatural Language Equivalent
/linguistic:lifecycleEnter the 5-phase pipeline"help me build an LLM for Yoruba"
/linguistic:exploreInvestigate a language/corpus interactively"what's in this CulturaX subset?"
/linguistic:decideForce a strategic decision"decide on the transfer source"
/linguistic:proposeGenerate a 5-phase plan"propose a roadmap for Khmer"
/linguistic:findingsList findings by severity"what are the issues"
/linguistic:reviewFull state + findings + history"where are we?"
/linguistic:statusOne-line phase summary"current phase?"
/linguistic:specCreate/edit a workspace spec"draft a tokenizer spec"
/linguistic:rollbackRestore prior workspace state"rollback to last good state"
/linguistic:helpShow this list"help"

Usage Pattern

Commands are used at decision points in the pipeline:

  1. Start: /linguistic:lifecycle or natural-language trigger
  2. Explore before committing: /linguistic:explore [language]
  3. Check status mid-session: /linguistic:status
  4. Force a pending decision: /linguistic:decide [topic]
  5. Generate a plan: /linguistic:propose
  6. Review all findings: /linguistic:findings
  7. Full briefing: /linguistic:review
  8. Document decisions: /linguistic:spec [type]
  9. Undo a wrong turn: /linguistic:rollback
Was this page helpful?
Edit on GitHub

Last updated on

On this page