Workshop: Rules and commands for Agentic IDE
Create effective rules and commands for modern Agentic IDEs (Cursor, Claude, Copilot) and start systematically managing agents.
Mentor - Viacheslav Koldovskyy, Competence Manager at SoftServe
Learn to create effective rules and commands for modern Agentic IDEs (Cursor, Claude, Copilot) and stop "just prompting" - start systematically managing agents. In two intensive sessions you will learn how to design rules for teams/projects, build a library of commands, configure safe "rails" for agents and get stable results from AI tools in real projects.
Content upgrade: we are adding Agent Skills – the newest and most practical tool for Agentic IDE today. We will build not just prompts, but a Rules + Commands + Skills system that can be scaled to teams and projects.
-
Understand how an agent IDE thinks:
- what are the types of prompts and rules (system, project, file-level, codebase rules);
- how do they affect the quality, cost and stability of the result;
- what are the typical mistakes when working with rules and commands
- Learn how to work with rules and commands in different tools, and what are the features for each tool (Cursor, Claude Code, GitHub Copilot).
-
Learn to build a framework of rules for a project:
- how to get a set of stable rules from requirements (PRD / tasks / user stories);
- how to formalize what is usually "in the head of the senior";
- how to use rules to teach the IDE to understand new versions of libraries and frameworks.
- Learn to solve typical cases and scenarios using rules and commands:
- scenarios for typical tasks (feature, bug fix, refactoring, performance, security)
- team templates for different roles: Dev, QA, DevOps, TL.
-
Understand approaches to security and quality:
- how to reduce hallucinations and "self-activity" of the agent due to rules
- how to describe license restrictions, quality standards and code style
- how to embed checks (tests, linters, review rules) into command instructions.
- Patterns for working with rules (Memory Bank, etc.).
In two practical sessions, you will:
set a task → apply rules and command → get code → validate → fix the improved rule.
- Practical skills in working with rules and commands in Cursor, Claude, Copilot.
- Ready-made instruction templates for individual work and for the team (separate roles of Dev, QA, DevOps, tech lead, architect).
- A framework for building rules for the project - from requirements and constraints to a clear set of instructions for agents.
- Understanding how to standardize work with AI tools in the team: so that the result is stable, reproducible and transparent in terms of costs.
- A clear idea of when "manual prompting" is enough and when it is worth moving to a full-fledged system of rules and commands.
Workshop outcome:
Format:
Materials: Video recording of the workshop (5 hours) and a presentation with a selection of materials
Language of the event: Ukrainian
After purchasing a ticket, refresh the page and you will see the video recording and all materials.
Language of the presentation: Ukrainian/English
Who will be interested?
- Developers of all levels and tech leads/architects, DevOps/SREs who want to stop “inventing prompts from scratch every time” and make interaction with agents predictable.
- Product/Project Managers, BAs, designers and other non-technical specialists – for prototyping and creating solutions independently without the participation of developers
- QA/Automation – for describing validation rules, test generation and standardization of quality approaches.
- Mentors, team leads and Engineering Management specialists who build standards for working with AI tools in teams and companies.
Program
Session 1: Theory of Rules and Commands for Agentic IDE
- Introduction: how “just prompting” differs from a structured system of rules and commands.
- Types of instructions in Agentic IDEs:
- system-level, project-level, and hierarchical rules
- rules by role (Dev, QA, Reviewer, Architect)
- Tool overview (Cursor, Claude Code, GitHub Copilot) and their features.
- How to turn requirements and constraints into rules:
- business constraints, security, licensing
- quality standards, architectural principles, code style
- Common mistakes and anti-patterns:
- rules that are too long or contradictory
- rules that are technically infeasible
- “magical” prompts without clear success criteria
- Q&A
Session 2: Practice — Building a System of Rules and Commands
- Preparing your IDE to work with rules:
- key IDE settings
- optimal project structure for efficient agent collaboration
- Practical case study:
- start with a simple PRD or set of requirements
- define a rule set (quality, security, style, architecture)
- create a command/script library for common tasks
- Working with multiple tools (Cursor, GitHub Copilot, Claude Code) and their specific characteristics.
- Completing the agent workflow:
- assigning a task → launching the command → receiving code
- validating results (tests, reviews, checks)
- updating rules and commands — how to evolve rule systems
- Checklist for participants:
- minimal set of rules for starting a new project
- structure of internal documentation with team rules
- how to onboard new developers with the rule system
- Q&A
The course concept and all course materials are the intellectual property of FVDAYS. Copying of course content and materials is possible only with the consent of FWDAYS.
Mentor: Viacheslav Koldovskyi
— Competence Manager at SoftServe
— 20+ years in IT, certified Google Cloud Professional Cloud Architect, nVidia Generative AI LLMs
— Ph.D / Candidate of Economic Sciences, Associate Professor, Head of the Gen AI Center at IT STEP University
— Founder of YouTube and Telegram channels Programming Mentor
— Active speaker: speaking at iForum, DOU Day, etc.
— Leader of the AI community at DOU, author of publications
— Has implemented projects with Gen AI and AI-generated code, which successfully working in production
— Advises companies on transforming SDLC processes using AI
Live course participant reviews: average score 9.5
Everyone who works with code, as Vyacheslav said, currently needs specialists who not only know a lot and know how to write code, but also those who know how to do it with modern tools
The workshop well shows the practical use of AI tools in large projects. For example, following the lecturer's recommendations, I was able to fix a bug in a project even with limited JavaScript knowledge, which well demonstrates the practical value of such an approach.
The workshop is simply great. Although, at first glance, it is focused on developers, as a tester I learned a lot of useful things for myself, which I have even applied to a working project. The presentation is very easy, the workshop was passed in one breath. I recommend it!
Event price
Presentation and materials
Access to the video recording of the workshop (5 hours)
Participants receive a 10% discount on participation in Fwdays conferences