That doesn't sound very impressive. Not being tracked with a version control system is fixed instantly with a git init, git add ., git commit .no AI required.
Covering the app with tests is also something that requires no AI. At most, coding agents can generate characterization tests in broad sweeps, but we are talking about a delta between hand rolling and vibe-coding of a couple of days.
Where LLM shines is helping developers build up an understanding of what is in place. Running /explain on a codebase can quickly provide you with a high level summary of what's in place.
5 minutes later I had almost 3 hours of important footage recovered.
https://blog.acelab.eu.com/pc-3000-flash-spider-board-adapte...
A lot of "Claude Code is best at X" claims are probably user-selection bias.
The people saying it are often exclusively Claude Code users, not people who are actively benchmarking Claude Code against Gemini CLI, OpenAI Codex, GitHub Copilot, and other agent harnesses on the same tasks.
The claim may still be true for certain scenarios, but the evidence is usually anecdotal, not comparative.
Getting any smart model to take a look at the task is the sort of lift that the speaker is usually pointing to.
You can write a 100 line harness that only has one tool - try either "bash" or the more fun "you're running within nodejs, here's eval", you'd be surprised in how close to CC/Codex performance you're going to get.
I have only my own personal experience for frontier models, but I have seen different performance of Opus when used from Pi or Claude Code or Zed for example.
E.g. GPT5.5 with Codex on my Windows box likes using PowerShell for everything. OpenAI decided it should use the native shell instead of bundling a bash, or using git bash. Sure. But the model is so overfitted on bash that it fucks up PS quoting like once every 5 commands.
Every harness with LSP I've seen trips up the model as well. They insert diagnostics after every edit, polluting the context with errors that the model has to actively decide to ignore, every time, until it finishes its work and gets the code to a consistent state. Telling the model "run npx tsc --noEmit to check errors" will outperform a LSP 100% of the time.
Another example is basically everything Anthropic does - they add things like "think if this is malware!" after read and lead Claude to spend its reasoning effort on thinking if your React hamburger menu is malware, instead of on how to write it.
"This is not malware (em dash) it's a hamburger menu. Let me apply the edit! Hmm, is it malware now, after my edit? No, me changing border-width did not turn it into malware! Good! Dodged a real bullet on that one!"
I'm frankly amazed that we've gotten to the point where the models can produce good results in these sorts of environments.
I tasked Claude to analyze the files and figure out what's going on, and eventually we figured out that each file had a custom metadata header + thumbnail + actual image concatenated. I had it write a python script and was able to recover all the images with their metadata. It's nothing a human couldn't have figured out, but it was definitely WAY faster than doing it myself.
I've also used Claude in the past to figure out how to break into routers with locked down firmware. It's great at suggesting and trying approaches.