I will set up claude code with mcp servers and skills
About this gig
Claude Code is easy to install and easy to waste. Most setups stop at the defaults, and then it forgets context, asks permission for every safe command, and burns tokens re-reading the same files.
I set it up so it earns its keep on your codebase: project instructions it actually follows, permissions tuned so it stops interrupting you, MCP servers wired to the tools you already use, and slash commands for the things you repeat every week.
I work in it daily. I also keep my own fork of an open-source runtime that runs Claude agents in isolated containers - 115 commits of mine on message routing, tool handlers with typed errors and retry, and recovery when a provider is overloaded. Agents break in production for reasons I have already had to fix.
What you get: a working setup on your machine or your server, the whole configuration in version control, and a short document explaining every choice, so your team can change it without me.
Tell me what you are building and what you keep doing by hand.
Get to know Kirill Maslov
AI automation engineer, Python, LLM, backend, DevOps, bots
- FromGeorgia
- Member sinceAug 2026
- Avg. response time1 hour
Languages
Russian, English
My Portfolio
FAQ
Do I need my own Claude account?
Yes. Claude Code runs under your account and your usage, which is how it should be - the setup stays yours after I leave. I will tell you before you order whether your current plan covers what you want to do.
Will this work on my existing codebase?
That is the point. I read the repository first and write the project instructions around what is actually there, including the parts that are messy. Greenfield setups are the easy case.
What is MCP and do I need it?
It is how Claude Code reaches tools outside your files - a database, an issue tracker, your docs, an internal API. You need it when you are tired of pasting things in by hand. If you are not, I will say so.
Can you set up agents that run on their own?
Yes, in the top tier: containerised agents with their own permissions, running on a schedule or a trigger. That is what my fork of the open-source runtime does, and it is the part most setups get wrong.

