Workshops: Agentic Coding Tools in Depth

Advanced Cursor Workshop: Project Rules, Background Agents & Bugbot

Cursor becomes far more useful once the whole team works from the same rules and lets agents do more than inline edits. This advanced workshop covers project rules that actually steer the agent, background agents that work on tickets in the cloud, Bugbot reviews on pull requests, MCP integrations and the team settings that hold it all together.

Who is this workshop for?

  • Developers who use Cursor daily and want more consistent agent results
  • Tech leads defining coding conventions for Cursor-based teams
  • Team admins responsible for privacy mode, models and settings
  • Maintainers who want automated review on GitHub pull requests
  • Teams deciding which work can move from the editor to background agents

What you will take away

  • Write scoped project rules with the right rule type, file patterns and examples, and retire outdated rule files.
  • Launch background agents on well-defined tickets, configure their environment and review the resulting branches efficiently.
  • Set up Bugbot with project-specific review guidance and build its findings into the team's review process.
  • Connect MCP servers for issue tracking, documentation or databases at project level with sensible permissions.
  • Define team-wide settings for privacy mode, model access and rules so new members are productive from day one.

Agenda

  1. Module 1

    Rules that steer the agent

    • Rule types: always applied, attached by file pattern, requested by the agent, manual
    • From one long rules file to small, scoped rules
    • Exercise: rules for your frontend, backend and test conventions
    • AGENTS.md and user rules next to project rules
  2. Module 2

    Agent workflows in the editor

    • Plan before edit: reviewing plans and checkpoints
    • Running several agents or models on one task and comparing results
    • When to stop the agent and take over
  3. Module 3

    Background agents

    • How they work: remote environment, own branch, asynchronous execution
    • Setting up environment, dependencies and secrets
    • Exercise: delegate two tickets and review the branches
  4. Module 4

    Bugbot and pull request review

    • Enabling Bugbot for repositories and adding review guidance
    • Triaging findings: fix in Cursor, discuss or dismiss
    • Working alongside human review and CI checks
  5. Module 5

    MCP and team settings

    • Project and user-level MCP configuration
    • Privacy mode, model choice and admin settings
    • An onboarding checklist for new team members

Why rules often fail

Many teams copy a long rules file from the internet and wonder why Cursor ignores half of it. Rules compete for context, vague wording is interpreted loosely, and rules that contradict the existing code lose. We audit your current rules, cut what does not help and rewrite the rest as short, testable instructions with examples from your own codebase.

Background agents need a different working style

An agent working in the cloud cannot ask you a quick question. Tasks must be self-contained, the environment must build and test the project without help, and the result arrives as a branch to review. Teams that treat background agents like a remote colleague with a clear brief and acceptance criteria get much better results.

Because background agents run on remote infrastructure with access to your repository, we also look at what this means for confidential code and your privacy settings.

Your setup, not a demo project

We work with your repositories and your Cursor team account, or with a sandbox copy if your code cannot be used. neooutput AI Code Engineers join you on-site or remotely, in English or German. You leave with revised rules and a checklist for team settings.

Frequently asked questions

Should we use .cursorrules or the rules folder?

The single .cursorrules file is the older approach. Scoped rules in the project rules folder are easier to maintain and only load when relevant. We help you migrate without losing useful content.

How do Cursor rules relate to AGENTS.md or CLAUDE.md?

Cursor can read an AGENTS.md file in addition to its own rules. If your team also uses Claude Code or OpenAI Codex, we show how to keep shared conventions in one place and tool-specific details in each tool's format.

What does a background agent see of our code?

It works on a copy of your repository in a remote environment run by Cursor. We walk through the relevant settings and help you decide which repositories and tasks are suitable; the data protection assessment stays with your team.

Workshop

Get in touch

Should Cursor follow your team's conventions without anyone having to remind it?

Tell us roughly how many people will join and pick dates that work – we will get back to you within 24 hours with a proposal for content, format and schedule.

Prefer to talk first?

15 minutes, free and without obligation – with our AI Trainers and AI Code Engineers.

Book a call for free (opens in new tab) +49 761 519 751 50 (24/7)

Request a training / workshop

Workshop: Advanced Cursor Workshop: Project Rules, Background Agents & Bugbot · 1 day

Dates that work for you (optional)

Suggestions about 2 weeks ahead – pick as many as you like

Reply usually within 24 hours · jocapps is ISO 27001 certified · Privacy