What Built-In Skills Does the OpenMAIC Agent System Provide?

OpenMAIC ships with 22 built-in skills stored in skills/agent-runtime that users invoke via the /skill-name syntax to automate curriculum design, content generation, and educational research workflows.

The OpenMAIC agent system includes a native skill framework that eliminates boilerplate setup for common educational AI tasks. These built-in skills are automatically discoverable and executable through the agent's read tool, with each skill defined by a self-contained SKILL.md file and optional JSON constraints. Users trigger complex pedagogical workflows by referencing the skill handle—the folder name—prefixed with a forward slash.

How OpenMAIC Built-In Skills Work

Built-in skills reside in the skills/agent-runtime directory, where each subdirectory represents a unique skill handle. When a user includes /skill-handle in a prompt, the system resolves this token through the composer logic and loads the corresponding SKILL.md definition.

The skill resolution pipeline is implemented in lib/workbench/composer-skills.ts, which handles:

• Token recognition – Parsing /skill-name patterns in user input
• Insertion and deletion – Managing skill references in the conversation context
• Rendering – Displaying skill handles with appropriate UI indicators

Unit tests in tests/workbench/composer-skills.test.ts verify that built-in skill tokens are correctly identified and processed, while tests/workbench/workbench-i18n.test.ts ensures skill names render properly across locales.

Complete Catalog of Built-In Skills

The OpenMAIC agent system provides 22 specialized skills organized by pedagogical function.

Curriculum Planning and Pedagogical Frameworks

These skills implement established educational methodologies and high-level planning workflows:

• curriculum-planner – High-level curriculum planning across grades and subjects, suitable for academic coordinators mapping multi-year instructional sequences.
• k12-core-literacy-planning – Subject-specific unit planning for mathematics, science, languages, and humanities within K-12 frameworks.
• understanding-by-design – Implements the Wiggins and McTighe "Understanding by Design" framework, focusing on backward design from learning outcomes.
• spiral-curriculum – Generates spiral-curriculum sequences where concepts are revisited with increasing complexity.
• learning-to-learn – Designs meta-learning activities that teach students cognitive strategies and self-regulated learning techniques.
• feynman-learning – Applies the Feynman Technique, guiding students to explain concepts in simple terms to expose knowledge gaps.

Instructional Delivery and Content Styles

These skills generate content using specific instructional modalities or clone existing stylistic patterns:

• lecture-style – Generates traditional lecture-style instructional narratives suitable for direct instruction models.
• workshop-style – Provides a workshop-style lesson design workflow emphasizing hands-on activities and collaborative learning.
• vocational – Generates vocational-training lesson plans tailored to technical and career education.
• teacher-style-clone – Analyzes and clones a specific teacher's stylistic tone, vocabulary, and pedagogical approach for lesson creation.
• style-clone – General-purpose style cloning for instructional text, applicable to institutional voice guidelines or specific genre requirements.
• page-clone – Clones page-level structural templates for consistent layout reuse across lesson materials.

Research and Validation

These skills support evidence-based content development and accuracy verification:

• deep-research – Conducts multi-step research and synthesis across sources, generating annotated bibliographies and literature reviews.
• deep-interactive – Facilitates interactive deep-dive tutoring sessions using Socratic questioning and adaptive scaffolding.
• fact-check – Performs automated fact-checking on supplied statements against trusted sources, flagging inaccuracies with corrections.

Presentation and Media Generation

These skills handle the creation of slide decks, staged flows, and multimedia imports:

• slide-dsl – Domain-specific language for authoring slide-based presentations with structural constraints.
• slide-craft – Crafting slide decks with content-aware defaults, including automatic layout suggestions and visual hierarchy optimization.
• pptx-import – Imports PowerPoint (.pptx) files into lesson components, extracting content while preserving structural relationships.
• stage-dsl – DSL for authoring staged instructional flows, defining prerequisite stages and progression gates.
• stage-design – Guided stage-by-stage lesson design wizard that breaks complex topics into scaffolded learning phases.

Authoring and Editing Utilities

Supporting tools for content refinement and skill development:

• pro-editing – Professional-grade editing and polishing of lesson drafts, including clarity improvements and pedagogical alignment checks.
• build-personal-skill – Boilerplate generator for authoring user-defined skills, providing templates and validation schemas for extending the skill library.

Skill Configuration and Constraints

Advanced skills may include optional outline-constraints.json files alongside their SKILL.md definitions. These JSON files specify curriculum standards alignments, stage dependency graphs, content length limits, and assessment rubric structures.

For example, skills/agent-runtime/curriculum-planner/outline-constraints.json likely defines the valid grade-level ranges and subject taxonomies available to that skill, while the narrative description and parameter schema reside in skills/agent-runtime/curriculum-planner/SKILL.md.

Summary

OpenMAIC delivers a comprehensive built-in skill ecosystem for educational AI development:

• 22 production-ready skills covering curriculum planning, content generation, research, and media creation
• Zero-configuration deployment via the skills/agent-runtime directory structure and native read tool integration
• Extensible architecture supporting custom skill development through the build-personal-skill template
• Robust composer integration with token recognition implemented in lib/workbench/composer-skills.ts and validated by comprehensive test suites

Frequently Asked Questions

How do I invoke a built-in skill in the OpenMAIC agent system?

Reference the skill by its folder name prefixed with a forward slash, such as /curriculum-planner or /fact-check. The composer logic in lib/workbench/composer-skills.ts detects this token and loads the corresponding SKILL.md definition from skills/agent-runtime/[skill-name]/.

Can I create custom skills in OpenMAIC?

Yes. Use the build-personal-skill built-in skill to generate boilerplate templates for custom skill development. Custom skills follow the same directory structure as built-in versions, requiring a SKILL.md file and optional outline-constraints.json for complex validation rules.

What is the difference between slide-dsl and slide-craft?

slide-dsl provides a domain-specific language for precise programmatic control over slide structure and content, while slide-craft offers a higher-level wizard interface with content-aware defaults and automatic layout suggestions for rapid deck creation.

Where are skill constraints defined in the OpenMAIC repository?

Optional constraints reside in outline-constraints.json files within each skill's skills/agent-runtime/[skill-name]/ directory. These files define validation rules for curriculum standards, stage dependencies, and content parameters, while the core skill description lives in SKILL.md.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}