Loadouts

Loadouts are bundled sets of content — rules, skills, agents, hooks, MCP configs — that apply as a unit. Think of them as named configuration profiles: instead of adding content piece by piece, you define a loadout once and apply everything in a single command.

Loadouts currently emit configuration for Claude Code only. Additional provider emitters are planned.

How loadouts fit in

Your library stores individual content items. Registries share content across teams. Loadouts sit on top of both — they bundle content from your library into apply-as-a-unit configurations.

Library (individual items) --> Loadouts (bundled profiles) --> Provider config (Claude Code)

Creating a loadout

Use the interactive creator to define a new loadout:

syllago loadout create <name>

This walks you through selecting which content items to include. The result is a named loadout stored in your syllago configuration.

Applying a loadout

The apply command has three modes controlled by flags:

# Preview what would change (default, no flags)
syllago loadout apply <name>

# Apply temporarily — auto-reverts when the session ends
syllago loadout apply <name> --try

# Apply permanently — persists until you explicitly remove it
syllago loadout apply <name> --keep

To target a specific provider:

syllago loadout apply <name> --keep --to <provider>

The default (no flags) is a dry run — it shows what files would be created or modified without touching anything. This lets you inspect the effect before committing.

Temporary vs. permanent

--try is designed for experimentation. Apply a loadout, test it during your coding session, and it automatically reverts when the session ends. Nothing to clean up.

--keep writes the configuration to disk permanently. To undo it, use syllago loadout remove.

Managing loadouts

# List all available loadouts
syllago loadout list

# Check what's currently active
syllago loadout status

# Remove the active loadout and restore original configuration
syllago loadout remove

Use cases

Per-project configurations. Define loadouts for different work contexts — a Python backend stack, a frontend stack, a security audit profile — and switch between them as needed.

Team onboarding. New team members apply the team’s standard loadout in one command instead of manually configuring rules, skills, and MCP servers.

Temporary experimentation. Use --try to test a new set of rules or skills without risk. If it doesn’t work out, it reverts automatically.

Collections overview — how library, registries, and loadouts work together
Library — your personal content store
Registries — shared team and community content
CLI reference: loadout commands — full flag and option details