本文へ移動
cccskills
無料GitHub で公開

simulink-create-course-activity

Create MATLAB Course Designer Simulink Exercise learning activities with starter and solution Simulink model files. Use when the user asks to create a Simulink activity, Simulink Exercise, starter model, solution model, model-based learning activity, or Course Designer-ready Simulink artifact. Uses MATLAB MCP Server tool calls and Simulink Agentic Toolkit guidance to create, inspect, edit, simulate, and validate model files.

インストール方法を見る

含まれるファイル(4)

  • SKILL.md7.6 KB
  • references/course-designer-artifacts.md2.7 KB
  • references/model-generation-workflow.md2.5 KB
  • references/validation-rules.md2.1 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Simulink Course Activity Generator

Create MATLAB Course Designer-ready Simulink Exercise learning activities. The skill must produce actual starter and solution model files, not only instructions. Use MATLAB MCP Server tool calls, following Simulink Agentic Toolkit guidance, to create, edit, inspect, and validate the models. The resulting activity folder should be ready for Course Designer web application import workflows, with course-designer-activity.json as the import-oriented metadata sidecar.

Input

The user may have provided a module objective or activity idea in their request. If it is missing or too vague, collect the missing activity inputs in Step 1.

Reference Files

Read these references from the references/ directory as needed:

  • references/model-generation-workflow.md — starter/solution model workflow
  • references/course-designer-artifacts.md — output folder and metadata rules
  • references/validation-rules.md — validation and review gates

Read references/model-generation-workflow.md and references/course-designer-artifacts.md before creating model files.

Required Tooling

Use Simulink Agentic Toolkit guidance for model construction and inspection. When editing models, follow the building-simulink-models workflow:

  1. Use model_read or model_overview before edits when a model exists.
  2. Use model_edit for structural edits and parameter configuration. Treat a status: partial result as a failure until every reported parameter rejection is resolved and the intended values are confirmed with model_query_params; a partial edit can leave a block at a silently wrong default that model_read and model_check do not surface.
  3. Use model_read, model_overview, and model_query_params to verify.
  4. Use MATLAB MCP calls such as evaluate_matlab_code, run_matlab_file, check_matlab_code, and detect_matlab_toolboxes for model setup scripts, toolbox checks, simulation smoke tests, and file existence checks.
  5. Use model_test for behavioral tests when a Gherkin validation spec is created, the testing workflow is available, and the model exposes at least one Inport; otherwise use the matlab.unittest fallback in references/validation-rules.md.

The building-simulink-models workflow starts with custom-library gates. For course activities built from base MATLAB and Simulink blocks, answer the custom-library question with "none" and keep the resulting .satk/ configuration in the course output folder, not in this skill package.

Do not rely on prose-only model descriptions for final artifacts.

Pipeline Overview

Step 1: Collect activity inputs and Course Designer placement
Step 2: Confirm products, toolboxes, and output folder
Step 3: Create the solution model with MCP tool calls
Step 4: Create the starter model from the solution model
Step 5: Validate solution and starter models
Step 6: Write Course Designer-ready activity metadata and review notes

Step 1: Activity Inputs

Collect only missing information:

  • Course title or course identifier
  • Module title and module order
  • Learning activity title and activity order
  • Module objective and observable learner outcome
  • Simulink concept: modeling, simulation, controls, physical modeling, verification, code generation, or another domain
  • Learner task and expected model behavior
  • Required MATLAB, Simulink, and toolbox products
  • Starter model scope: what should be missing, incomplete, or configurable
  • Solution model scope: complete expected behavior
  • Validation requirements: simulation output, signal behavior, parameter values, model structure, or manual rubric
  • Output directory

Default to a Course Designer learning activity type of Simulink Exercise.

Step 2: Product and Output Checks

Use detect_matlab_toolboxes when available to confirm MATLAB, Simulink, and required toolbox access. If a needed toolbox is missing, stop before creating a misleading model and report the blocked requirement.

Create or use this output folder:

simulink-activities/module-NN-activity-MM-[slug]/

All generated file names must be stable, lowercase, and Course Designer import friendly.

Step 3: Create Solution Model

Use references/model-generation-workflow.md.

Create the complete solution model first. The solution model is the instructor reference and the source used to derive the starter model.

Required actions:

  1. Create or open the model using MATLAB MCP calls.
  2. Use model_edit for blocks, connections, parameters, model configuration, and subsystem structure.
  3. Use model_read or model_overview to verify topology.
  4. Save the model as:
[activity-slug]_solution.slx

Step 4: Create Starter Model

Create the starter model from the validated solution model. Remove, mask, disable, parameterize, or replace only the parts learners are expected to complete.

Required actions:

  1. Copy or save the solution model as:
[activity-slug]_starter.slx
  1. Use model_edit to create the intended learner gaps.
  2. Use model_read or model_overview to confirm the starter model still opens and contains the expected scaffold.
  3. Do not leave broken model references, missing files, or unresolved variables unless they are explicitly part of the learner task.

Step 5: Validate Models

Use references/validation-rules.md.

At minimum:

  • Confirm both .slx files exist.
  • Open or inspect both models through MCP calls.
  • Run a simulation smoke test on the solution model when simulation is part of the activity.
  • Confirm the starter model contains the intended learner gaps.
  • Confirm required variables are supplied through a model workspace, data dictionary, or init script included with the activity.
  • Write validation results to validation-report.md.

When behavioral validation is needed, create a Gherkin validation spec and run model_test when the testing workflow is available and the model has at least one Inport; for self-contained stimulus models, use the matlab.unittest fallback in references/validation-rules.md instead.

Step 6: Course Designer Activity Artifacts

Use references/course-designer-artifacts.md.

Write these files in the activity folder:

  • [activity-slug]_starter.slx
  • [activity-slug]_solution.slx
  • activity.md
  • instructor-notes.md
  • validation-report.md
  • course-designer-activity.json
  • Optional init scripts, data files, requirements files, or validation specs

The course-designer-activity.json file is the import-oriented sidecar for future Course Designer API work. It must include course, module, learning activity, model file, product, and validation metadata.

Output Rules

  • The final answer must list the starter model path, solution model path, metadata path, and validation report path.
  • Do not claim the models are ready unless the validation gates passed.
  • If validation is partial, state exactly which checks were deferred.
  • Keep starter and solution models separate; do not overwrite one with the other.
  • Preserve the Course Designer hierarchy: course -> module -> learning activity.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Use this skill when the user wants to author, design, scope, or refine an Agent Skill (a SKILL.md file). Trigger phrases include "build a new skill", "design an agent skill", "scope a SKILL.md", "how should I structure this skill", "write a skill for X", "my skill isn't working well", or any request to improve an existing SKILL.md. Walks the user through an empirical, test-first process — probe the agent for real failures, design only for genuine knowledge gaps, iterate against runnable examples, and verify across models.

日本語の概要は準備中です。原文の説明を表示しています。

matlab/agent-skills-playground1842026年10月10日 更新

Deploy AI models to embedded hardware using MathWorks tools (MATLAB, Simulink, Embedded Coder). Covers two workflow patterns: (1) MathWorks-native or 3P-imported models rebuilt as dlnetwork for lean hardware (Cortex-M, DSP), (2) direct C/C++ code generation from PyTorch and LiteRT models for high-performance hardware (Cortex-A, x86, GPU). Trigger when: user wants to deploy AI to embedded targets; generate C/CUDA from neural networks; compress AI models for MCU/DSP; integrate AI in Simulink for system-level simulation; import PyTorch/ONNX/TensorFlow models for embedded deployment; optimize AI for resource-constrained hardware; or use loadPyTorchExportedProgram, importNetworkFromPyTorch, dlquantizer, exportNetworkToSimulink, or Embedded Coder with AI models.

日本語の概要は準備中です。原文の説明を表示しています。

matlab/agent-skills-playground1842026年10月10日 更新

Use when a learner asks for help with MATLAB homework, labs, projects, graded assignments, take-home exams, quizzes, or any programming task where academic integrity, course policy, or instructor constraints may limit direct solutions. Use to provide policy-aware hints, conceptual coaching, partial feedback, and assignment-safe MATLAB tutoring.

日本語の概要は準備中です。原文の説明を表示しています。

matlab/agent-skills-playground1842026年10月10日 更新

Use when tutoring a learner through MATLAB debugging, error interpretation, failed tests, incorrect outputs, array-shape problems, indexing mistakes, function argument issues, or code repair practice. Use for guided debugging sessions, debugging drills, teach-the-agent critique, and evidence-based MATLAB troubleshooting.

日本語の概要は準備中です。原文の説明を表示しています。

matlab/agent-skills-playground1842026年10月10日 更新

Use when an AI tutor session concerns MATLAB programming concepts, MATLAB syntax, MATLAB errors, MATLAB code style, MATLAB projects, or MATLAB toolbox workflows.

日本語の概要は準備中です。原文の説明を表示しています。

matlab/agent-skills-playground1842026年10月10日 更新

Use when an instructor wants to create, interview for, configure, install, update, or review a course AI-use policy for MATLAB AI tutoring. Produces an AI-POLICY.md file for LMS sharing and local tutoring-session enforcement by assignment guardrails.

日本語の概要は準備中です。原文の説明を表示しています。

matlab/agent-skills-playground1842026年10月10日 更新

matlab のスキルをすべて見る

このスキルの問題を報告する