Workshops: Agentic Coding Tools in Depth

MCP Servers in Daily Development: Hands-On Training for Dev Teams

MCP servers connect coding agents to the systems developers use all day: issue trackers, error monitoring, databases, documentation and more. This training is about using them, not building them: which servers are worth it, how to configure them in Claude Code, Cursor, GitHub Copilot or Codex, and how to keep permissions tight so convenience does not turn into a security gap.

Who is this workshop for?

  • Developers who want their coding agent to reach their everyday systems
  • Tech leads who want shared, version-controlled MCP configurations for their team
  • Support and on-call engineers who investigate errors and incidents
  • Teams with Jira, Linear, GitHub Issues, Sentry, Postgres or similar systems
  • Security contacts who need to understand what MCP connections expose

What you will take away

  • Choose the MCP servers that genuinely improve your workflow and skip those that only add noise.
  • Configure MCP servers at user and project level in Claude Code, Cursor, VS Code with GitHub Copilot and OpenAI Codex.
  • Run end-to-end workflows: from ticket to pull request, from production error to fix, from schema question to migration.
  • Apply least privilege with read-only tokens, narrowly scoped OAuth access and suitable tool approval settings.
  • Keep the context window lean by enabling only the tools a task needs.

Agenda

  1. Module 1

    MCP from the user's side

    • What a server provides: tools, resources and prompts
    • Local and remote servers, sign-in via OAuth or tokens
    • Vendor servers from GitHub, Atlassian, Sentry or Linear versus community servers
  2. Module 2

    Setting up your toolbox

    • Configuration in Claude Code, Cursor, VS Code and Codex
    • Project configuration in the repository or personal settings
    • Exercise: connect your issue tracker and code host
  3. Module 3

    Workflows that pay off

    • Ticket to pull request with full issue context
    • Error monitoring: from stack trace to reproduction and fix
    • Database servers for schema exploration, read-only by default
    • Documentation servers for current library docs
  4. Module 4

    Permissions and risks

    • Read-only first, write access only with a reason
    • Prompt injection via ticket text, error messages and web content
    • Approval prompts: when to confirm every call and when not
  5. Module 5

    Keeping it manageable

    • Too many tools: context consumption and wrong tool choices
    • Shared team configuration and onboarding of new colleagues
    • Tracking server provenance and updates

From copy and paste to connected context

Without MCP, developers paste ticket text, stack traces and query results into the chat by hand. With well-chosen servers the agent fetches the context itself, sees the current state and links its work back to the ticket. That saves time, but only while the selection stays focused.

We start with two or three servers that match your daily pain points and expand step by step, instead of connecting everything on day one.

Safety rules we practise

Every MCP connection widens what an agent can see and do. So we practise a short set of rules participants apply to every new server.

  • Dedicated tokens with minimal scopes, never personal admin tokens
  • Databases only as development or anonymised copies
  • All returned content treated as untrusted input
  • Servers from the system's vendor or vetted sources preferred

Scope and format

Building your own MCP servers and company-wide governance of third-party servers are separate workshops, as is browser automation with Playwright MCP. This day focuses on productive, safe use in the team's daily work, hands-on with your tools, remote or on-site, in English or German.

Frequently asked questions

Which MCP servers should we start with?

Usually the issue tracker, the code host and error monitoring, because they show up in almost every task. Databases and documentation follow once the team handles permissions and approvals confidently.

Can we share MCP configurations across the team?

Yes. Several tools support project configuration files that live in the repository, while credentials stay in each developer's environment or a secrets manager. We set up such a configuration together.

Is it safe to use MCP servers?

As safe as their configuration and provenance. Narrowly scoped tokens, read-only defaults, approval prompts and vetted sources reduce the risk considerably, and we show you how to apply them.

Workshop

Get in touch

Which systems should your coding agent reach directly first?

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: MCP Servers in Daily Development: Hands-On Training for Dev Teams · 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