Engineering guidance
This section collects the rules that keep LocalGPT maintainable after the architecture has become interesting enough to bite back.
Working principle
Use the strongest available evidence and label weaker evidence honestly:
real build/test > parser/static validation > focused inspection > assumption
A capability gap should report the requested outcome, missing dependency or source, evidence inspected, safe next step, and whether owner-side tooling or approval is required.
Open work
Current owner-side verification tasks live on the maintenance status page. Historical task lists and pass manifests remain under docs/internal-notes and are excluded from the reader site.