Repository navigation
.pr_agent_accepted_suggestions
qodo-merge-bot edited this page Nov 21, 2026
·
2 revisions
| PR 67 (2026-11-21) |
[correctness] Windows CLI error helper fails to parse
Windows CLI error helper fails to parse
`Show-UserIssue` has a broken `$Code -match` condition that begins with `'^[a-z_]+`, splices a duplicated `Write-Host` block into the statement, and leaves the friendly title and next-action output duplicated. When a Windows entrypoint dot-sources `scripts/user-errors.ps1`, the malformed statement can stop the workflow before any stage runs; in TUI mode, failure to emit `EPLUS_ISSUE:` also leaves `cli_issue_outcome` without the specific guidance and step-two-or-later payment-history warning.[correctness] Unexpected start failures lose their error detail
Unexpected start failures lose their error detail
The `except Exception as error` handler in `run_stage` replaces the exception type and message with a generic Outcome without recording a sanitized cause in the UI or Debug log. When `command_for_step` raises an actionable start error or `Popen` fails, the status points users to technical details that contain no explanation of the failure.[correctness] Button failures miss payment checks
Button failures miss payment checks
`Get-UserIssue` adds a second `subscribe_button_not_found` case, but its earlier case returns before the new guidance can run. When a checkout button is not found, callers still receive the old screen-check message rather than the added instruction to check pricing and payment history.[correctness] Malformed protocol lines reach the visible log unchanged
Malformed protocol lines reach the visible log unchanged
`readable_issue_line` returns any `EPLUS_ISSUE:` line that does not fully match `[a-z_]+` unchanged, so a marker followed by an email address is passed through as-is. `_received` then writes that raw line to the visible RichLog and the Debug log, where it is not masked. The added test asserts this passthrough, so the behaviour is intentional but it exposes account data and the raw protocol marker.| PR 63 (2026-11-21) |