| # Fix-ClaudeDesktopMSIX.ps1 | |
| # | |
| # Diagnose + repair ladder for Claude Desktop (Windows MSIX) stuck in | |
| # "Modified, NeedsRemediation" / refusing to launch. | |
| # | |
| # Field-tested 2026-08-24 on Windows 11 (build 26200) while debugging this class: | |
| # https://github.com/anthropics/claude-code/issues/84333 (and siblings) | |
| # | |
| # Key facts this script encodes (all reproduced live): | |
| # 1. Running `Add-AppxPackage -Register` on a HEALTHY package is a CORRUPTION |
One tiny SessionStart hook that tells Claude (and you) when a session started in the wrong folder — before you lose a day of context.
Claude Code keys three things to the directory a session starts in:
~/.claude/projects/<encoded-cwd>/memory/
| #!/usr/bin/env python3 | |
| """claude_task_audit.py - audit Claude Desktop scheduled tasks: which ones are | |
| reusable, and which ones will stop and ask for permission on their next run. | |
| Why this exists | |
| --------------- | |
| Scheduled ("routine") tasks each carry their OWN permission mode, and the desktop | |
| UI is the only place that shows it. If a task falls back to a prompting mode, you | |
| find out hours later, from a dialog nobody clicked. Two facts make that auditable | |
| and largely avoidable, both verified by measurement (2026-08-24, Windows hub + |
Your Claude Desktop routines vanished from the Routines list — after an app update, after signing in with a different account, or for no visible reason at all. This finds them and puts them back.
They are almost certainly not deleted. The app splits a routine across two places:
| what | where | survives an app update? |
|---|
| #!/usr/bin/env python3 | |
| """api_wait_gaps.py - measure API-wait stalls across a whole Claude Code transcript corpus. | |
| A "stall" here = wall-clock time in which the RUNTIME owed the next event and no bytes | |
| arrived. Only two record pairs qualify: | |
| tool_result -> assistant (tool came back, model has not spoken yet) | |
| assistant -> assistant (mid-stream continuation) | |
| Why the restriction matters: a naive "every gap >= 60s" scan over a corpus is dominated by |
| """Verify headroom v2.0.8's MessageDeduper against a Windows hub log tree. | |
| Simulates the shipped O(1) rule (drop a usage record whose message.id equals the | |
| PREVIOUS record's message.id) and measures what inflation survives it, then | |
| classifies the survivors: | |
| A REPLAY - the identical record (same uuid) is written again later in the file | |
| B PARTIAL - one message.id carries DIFFERENT usage payloads (streaming snapshots) | |
| Ground truth for "how many real messages" = distinct uuid, since uuid is unique | |
| per written record and a replay repeats it verbatim. | |
| 0 LLM tokens, read-only. |