Skip to content

Module Management

Lola modules (LoLaS) are portable packages of skills and context for AI assistants. A module can contain a single skill, multiple skills, or a full AI Context Module with commands, agents, and MCP servers.

Adding Modules

# From a git repository
lola mod add https://github.com/user/my-skills.git

# From a local folder
lola mod add ./my-local-skills

# From a zip or tar file
lola mod add ~/Downloads/skills.zip

# From a monorepo with custom content directory
lola mod add https://github.com/company/monorepo.git --module-content=packages/lola-skills

# From a flat repository (use root directory)
lola mod add https://github.com/user/flat-repo.git --module-content=/

When adding from a local folder that is inside a git repository, Lola honors the repository's .gitignore and skips ignored files. Common development directories such as .git/, .venv/, node_modules/, and Lola's own install paths are also excluded from the copied module.

Managing Modules

# List registered modules
lola mod ls

# Show module details
lola mod info my-skills

# Update module from source
lola mod update my-skills

# Remove a module
lola mod rm my-skills

Module Structure

Lola supports two module types. See Skills and Context Modules for full definitions.

Agent Skills

Agent Skills come in two forms:

Standalone — a single skill following the agentskills.io standard:

my-skill/
  SKILL.md           # Required
  scripts/           # Optional

Skill Pack — multiple related skills grouped together:

my-skills/
  skills/
    skill-a/
      SKILL.md
    skill-b/
      SKILL.md

Complete module with instructions, skills, commands, and agents:

my-module/
  module/
    AGENTS.md        # Module-level instructions
    skills/
      skill-a/
        SKILL.md
        scripts/
    commands/
      review.md
    agents/
      helper.md

Appending Context References

When an AI Context Module has files that reference each other, for example an AGENTS.md that says:

Follow the coding conventions in `context/conventions.md`
Run the setup script at `scripts/bootstrap.sh`

You can use --append-context so the agent reads the original file where these paths resolve naturally:

lola install my-module --append-context module/AGENTS.md

This appends a reference in the target assistant's instruction file (CLAUDE.md, GEMINI.md, AGENTS.md, etc.) pointing to the file inside .lola/modules/:

Read the module context from `.lola/modules/my-module/module/AGENTS.md`

Without the flag, the default behavior copies content verbatim, which works well for modules without relative path references.

:exclamation: NOTE Using --append-context adds an extra layer of file reading for the agent - it first reads the assistant's instruction file, then follows the reference to read the appended context file. For best performance, we recommend structuring your module to work with the default installation when possible, using --append-context only when your module requires relative path references between context files.

Content Path Detection

Lola auto-detects where module content lives. It checks for a module/ subdirectory first, then falls back to the repository root. Override with --module-content:

lola mod add https://github.com/company/monorepo.git --module-content=packages/ai-tools