Custom MCP Skills

Overview

The Webiny MCP server ships with a set of built-in skills that cover standard development patterns — content models, lifecycle events, dependency injection, GraphQL schemas, infrastructure, and more. For most projects, these built-in skills are all you need.

However, as your project grows, you may want to extend the MCP server with project-specific knowledge — internal conventions, custom abstractions, domain-specific patterns, or team-specific workflows. You can do this by adding custom skills folders or replacing the built-in skills entirely.

Adding Extra Skills

Use the --additional-skills flag to add a folder of custom skills on top of the built-in ones. Update the MCP server entry in your agent’s config file:

You can repeat the flag to add multiple folders:

When a skill with the same name exists in multiple folders, the one in the highest-priority folder wins. Priority is determined by order: folders listed later on the command line take precedence over earlier ones, and all additional folders take precedence over the built-in skills.

Replacing Built-in Skills

Use the --skills flag to replace the built-in skills folder entirely:

When you use --skills, only skills found in the specified folder are served. The built-in skills are ignored completely.

Writing a Skill

Each skill is a folder containing a single SKILL.md file. The file must start with YAML front-matter:

Front-matter fields:

• name (required) — unique identifier for the skill, used as the topic argument in get_webiny_skill
• description (required) — shown in the skill catalog; helps the agent decide which skill to load
• context (optional) — groups skills in the catalog; defaults to webiny-extensions

Place the SKILL.md file inside a descriptively named folder:

Paths passed to --skills and --additional-skills are resolved relative to the current working directory (your project root).

When to Use Custom Skills

Custom skills are useful when your project has patterns or conventions that go beyond what the built-in skills cover. Some examples:

• Domain-specific models — document your project’s content model relationships, naming conventions, and field patterns so AI generates models that fit your domain
• Internal APIs — describe custom GraphQL schemas or third-party integrations your team has built, so AI knows how to use them
• Team conventions — codify coding standards, folder organization rules, or review checklists that your team follows
• Migration guides — provide step-by-step instructions for common data migrations or schema changes specific to your project