Command Reference

In the current Cortex model, slash commands are not best understood as a separate hand-curated product surface. They are primarily generated or derived from the skill library.

How Commands Work

Two implementation details matter:

  1. cortex install link generates ~/.claude/commands/ aliases from skills/
  2. command metadata is discovered from SKILL.md files, with the command either derived from the skill path or overridden by frontmatter

That means the most accurate source of truth for a command is usually the skill that backs it.

Practical Mental Model

Think of it this way:

  • skills are the canonical reusable workflows
  • slash commands are one invocation surface for those skills
  • the installed commands/ directory is generated from the skill set rather than maintained as an entirely separate authored catalog

Common Patterns

Default namespace

Flat skills typically map into a ctx namespace:

skills/systematic-debugging/SKILL.md -> /ctx:systematic-debugging

Nested namespaces

Nested skills can map into their own namespace:

skills/collaboration/writing-plans/SKILL.md -> /collaboration:writing-plans

Explicit override

A skill can also define an explicit command in frontmatter:

command: /ctx:systematic-debugging

What To Read Instead Of Treating This As A Static Catalog

Notes

  • Avoid relying on older fixed command counts in site/.
  • If command behavior seems unclear, inspect the backing skill first.
  • If the generated aliases differ from older docs, trust the installed skill surface and current CLI behavior rather than the legacy command-count copy.