Skills
A skill is a set of instructions that defines a reproducible engineering procedure: when to apply it, what data is required, what to do, how to verify the result, and what to pass to the user. The skill may include reference materials and utility scripts.
The tool performs a separate action, for example, runs the Julia code. The skill determines the sequence of these actions for a specific task, for example, building RF component or checking the connection with RITM. The skill does not relearn the Large Language Model (LLM) and does not replace a license, library, or access to hardware.
How to use skills
-
Describe the task, the initial data, and the expected outcome.
-
The agent checks the environment and gets acquainted with the available skills through
skill_listor from the informationengee_start. -
Uploads a suitable skill via
skill_loadfor example withskill_path: "engee_model". -
Downloads specific help files as needed.
-
Performs the procedure, checks the result, and reports what is confirmed by the calculation and what remains a limitation.
You can specify the skill explicitly:
Используй навык engee-rf-modeling. По приложенной документации создай модель указанного РЧ-компонента и отчет о допустимых уровнях детализации. Сначала проверь достаточность исходных данных.
Catalog of skills
Below is a catalog of 26 Engee MIND skills. The composition may vary; user skills are added separately. Technical identifiers are saved verbatim, including the spelling developement.
| Identifier | Appointment |
|---|---|
|
Creating, calculating, diagnosing, and saving models Engee |
|
Modeling of a single RF component based on documentation and data in EngeeRF |
|
Structural analysis |
|
Developing a custom block C Function |
|
Block development Engee Function with a custom code |
|
Preparation and launch Engee.Integrations, connecting a local client |
|
Passive status check Engee.Integrations |
|
View, create, and delete instances of RITM machines in the list |
|
Checking the availability and system operations of the selected machine. |
|
Assembling, deploying, starting and stopping models on the RITM |
|
Versions and updates of the RITM firmware and support package |
|
Getting simulation results and files from the RITM into the session Engee |
|
Organization of development Genie App and readiness monitoring |
|
Preparing the repository for Genie App development |
|
Designing the application interface |
|
Implementation of the client part |
|
Application Server Logic |
|
Checking contracts of functions and objects Engee for the application |
|
Investigating the behavior of existing MATLAB code |
|
Modular, API, and static checks |
|
End-to-end application verification in the browser |
|
Git Operations and Genie App Deployment |
|
Creating and updating user skills Engee MIND |
|
Julia’s service load of subject skills resources |
|
Skill format template; not an independent engineering scenario |
Sources and priority
Skills can come from the server’s built-in kit (builtin), a general updatable set (volume) or the user environment (pod). If the applicable versions match, priority applies.:
pod → volume → builtin
Therefore, a locally installed user skill can change the procedure that the agent performs. To understand which text is used, ask to show the source and path of the selected skill, and then upload it. Downloading all the skills at once is not required and consumes the LLM context.
If there is no suitable ready-made skill in the task, the agent can perform operations with the Engee MIND tools according to your requirements. Any mention of the script should not be considered proof of the existence of a separate automatic converter or a special tool.