Find where your AI agents make errors, and why.
I review one workflow in which AI agents act for your team: where their errors happen, why, and how to catch them. You get a written analysis, not a promise.
Write to Ivan → [address]When to call.
- Your team uses AI coding agents every day, and nobody can say which of their mistakes review catches and which get through.
- A reviewer keeps rejecting agent work, and you cannot tell whether the agent erred or the standard was never written down.
- An agent did something costly, and you want to know what it knew at the time and what else it could have done.
- You are about to give agents more of a process and want to know where the checks belong.
What you get: an agent error review.
A written report on one workflow you choose, in three parts.
- The points where an agent's action becomes final and hard to undo.
- Each error found in your records, with the agent, the standard it missed, what it knew at that moment and what it could have done instead; and why each kind happened, as far as the evidence shows and where it stops.
- The checks that would catch each kind before it becomes final: in review, in tests, or in the process.
- Scope and duration: [Ivan sets them]. Price: [fixed, quoted in writing after the first call — Ivan decides, ask 3].
What it does not promise.
- It does not predict your agents' next errors. Whether agent errors can be predicted is the question my research asks, and it is not yet answered.
- It does not depend on your agents being bad at anything. If they make few errors, the report says so.
- It is an investigation. The result is what the evidence shows, including "your workflow already catches what matters".
How it works.
- Write to me with a few lines about your team and the workflow.
- [A free first call — Ivan decides, ask 3]: you describe the workflow, I say whether the review fits, and I check the work against my existing commitments before accepting it.
- A written scope [with a fixed price — ask 3]. You share the records you can: agent sessions, reviews, tickets, commits.
- I judge each candidate error against a standard written down before the analysis starts, write up what I find, and we go through it together. You keep the report.
The research behind it is public.
The review comes from an open research project on the philosophy of errors: what errors are, how failures emerge in complex systems, and whether dangerous combinations of conditions can be recognised before a major failure. Its current study covers AI coding agents. Drafts are published at philosophyoferrors.org, each marked with what is and is not established.
Who does the work.
Ivan Iufriakov, in Charlotte, North Carolina. [Two sentences of background in Ivan's words; any figure in them confirmed by him.] Mix Tools LLC is his company. Client material stays private.
Write to me.
Describe the workflow and the agents in it in a few lines. [I read and answer every message myself — Ivan confirms, ask 3.]
Email: [address]