An agent that has silently lost its tools does not report that it lost its tools. It answers anyway, with the same fluency as before.
Arthur VinbergHead of DevOps & IT · Leader, 
Ionix
Context. From the post's account of a sandbox failure in which the review agent could not run any commands.
Post: We built an AI PR reviewer. The hard part was teaching it to say nothing., Ionix blog, August 13, 2026, at Two failures that taught us more than the design did.
Arthur Vinberg, Head of DevOps & IT. Wrote the post about a system his team built and runs; an insider describing his own company's tool, with no figures to audit.
Checked against the source by a first reader, then confirmed independently by a second reader who fetched the source without seeing the first reader's notes (Independent review agent (Claude Sonnet), blind second pass from passage text and citation only, no repository access, 2026-11-08).
Found verbatim in the section 'Two failures that taught us more than the design did'. The post is bylined Arthur Vinberg, Head of DevOps & IT, and is written in first person plural. Only tags, entities and curly quotes were normalized.
Arthur Vinberg (Ionix), “We built an AI PR reviewer. The hard part was teaching it to say nothing.”, Ionix blog, 2026-08-13, at Two failures that taught us more than the design did. https://www.ionix.io/blog/we-built-an-ai-pr-reviewer/ Archived at https://the-ai-record.com/passages/ionix-agent-silently-lost-tools-answers-anyway-2026
The first was cost. We were paying more for it than the results justified, and the obvious replacements wanted the same kind of money.
Arthur VinbergHead of DevOps & IT · Leader, Ionix
Context. The post's author explains why Ionix dropped Rovo, an AI review tool, before building its own reviewer; “it” in the passage is Rovo.
Post: We built an AI PR reviewer. The hard part was teaching it to say nothing., Ionix blog, August 13, 2026, at Why we built it instead of buying it.
Most days the entire output is one sentence saying the change looks fine, which is what we built it to do. On the days it leaves an inline comment, someone reads it.
Arthur VinbergHead of DevOps & IT · Leader, Ionix
Post: We built an AI PR reviewer. The hard part was teaching it to say nothing., Ionix blog, August 13, 2026, at If you’re building one of these.
The second is permission to find nothing. The prompt says outright that returning zero findings is the correct and expected outcome for a clean PR, and that findings must never be manufactured to look thorough.
Arthur VinbergHead of DevOps & IT · Leader, Ionix
Post: We built an AI PR reviewer. The hard part was teaching it to say nothing., Ionix blog, August 13, 2026, at The noise budget.
So the reviewer is deliberately a different stack from our authoring tools: OpenAI’s Codex CLI, pointed at a model deployment we host on Azure AI Foundry.
Arthur VinbergHead of DevOps & IT · Leader, Ionix
Post: We built an AI PR reviewer. The hard part was teaching it to say nothing., Ionix blog, August 13, 2026, at A reviewer that didn’t write the code.
A lot of our code is now written with an AI assistant in the loop. The first thing we tried was the obvious thing: ask the same Claude session that wrote the change to review it. That does not work. It missed critical bugs, and it missed them confidently, because it had already reasoned its way to believing the code was correct, and re-reading its own work only confirmed the conclusion.
Arthur VinbergHead of DevOps & IT · Leader, Ionix
Post: We built an AI PR reviewer. The hard part was teaching it to say nothing., Ionix blog, August 13, 2026, at A reviewer that didn’t write the code.
Something wrong? Report a correction. JSON: /api/v1/passages/ionix-agent-silently-lost-tools-answers-anyway-2026.json