I will set up a universal ai coding agent harness


About this gig
Your team uses AI coding agents in the terminal, in the IDE or in the cloud, but the results are uneven. Agents ignore your conventions, break working code, and nobody knows what was checked. I set up a harness: the rules and checks that make agent work predictable.
One universal harness: the same rules, roles and checks work across all major agent platforms, so you can switch or mix agents without starting over.
What you get:
- Portable rule files and agent roles for your repository
- Skills and hooks for repeated tasks
- Pre-merge checks every agent change must pass
- An independent validator agent (Premium)
- A written guide for your team
Why me:
20+ years of hands-on programming and 20 years as CTO of a 180-person software company. I built its delivery process: specification, review, QA, handover. A harness is the same process, built for AI agents. I run multi-agent teams on my own production systems; one of them has 330+ tracked tasks.
I work in your repository through Git. Everything in writing. Before ordering, tell me your stack and which agents you use.
Get to know Oleg S
Systems Architect, custom business systems and AI agents
- FromBulgaria
- Member sinceNov 2026
- Avg. response time1 hour
Languages
English, Ukrainian, Russian, Bulgarian
My Portfolio
FAQ
Which agents do you support?
All major AI coding agents: in the terminal, in the IDE and in the cloud. The rules use portable formats, so one harness serves several agents.
What access do you need?
Access to the repository or a copy of it. No production secrets or passwords.
Will it work with our CI?
Yes. Checks run in GitHub Actions or GitLab CI.
Can we switch agents later?
Yes. All rules, skills and checks are files in your repository. They stay useful when you change or add an agent.
