Why
The v0.5.2 heuristic retune for is in good shape and has focused heuristic tests, but current coverage stops at the core scoring/assessment logic.
We do not yet have an integration-style test that exercises the real CLI/store flow end to end.
Gap
Current tests cover:
- recent broad weakly anchored nodes can be flagged
- specific well-anchored nodes are not casually flagged
- age increases severity but is not the admission gate
Current tests do not yet prove:
- dry-run output over a real temp DB
- apply path writes hygiene state as expected
- clear path removes hygiene state cleanly
- CLI argument handling stays wired correctly
Goal
Add one or two lightweight integration tests around the real script entry points and temp DB usage so future heuristic refactors cannot silently break the maintenance flow.
Notes
Keep this narrow. This is about CLI/store integration confidence for the hygiene maintenance script, not a redesign of the heuristic or store model.
Why
The v0.5.2 heuristic retune for is in good shape and has focused heuristic tests, but current coverage stops at the core scoring/assessment logic.
We do not yet have an integration-style test that exercises the real CLI/store flow end to end.
Gap
Current tests cover:
Current tests do not yet prove:
Goal
Add one or two lightweight integration tests around the real script entry points and temp DB usage so future heuristic refactors cannot silently break the maintenance flow.
Notes
Keep this narrow. This is about CLI/store integration confidence for the hygiene maintenance script, not a redesign of the heuristic or store model.