diagnose-fail-closed-upgrade

Stuck engine upgrades and git pull failures are resolved by triggering preflight checks to isolate the exact failure point, then validating system state against guard logic to determine if a code-level fix or configuration cleanup is required.

A preflight check blocks an engine hop or pull due to a stale lock file because safety guards detect configuration drift or stray core.lock integrity issues, requiring root-cause analysis to distinguish transient locks from genuine guard errors.

Troubleshoot git remote normalization issues by running a diagnostic reproduction of the preflight gate to analyze the output, which identifies whether mismatched remotes require a configuration cleanup or a script-level guard fix.

Yes, you need bash, python3, git, and jq installed, because the diagnostic scripts depend on these tools to reproduce preflight checks and evaluate guard logic against the current system state.

The best way to distinguish a code-level guard error from transient configuration drift is to safely trigger preflight checks and validate the system state against guard logic, isolating the failure point without entering infinite loops.

A key limitation is that standard troubleshooting can trigger infinite loops during preflight checks, so you must use a controlled one-shot reproduction to safely isolate the failure point and avoid recursive guard failures.