Problem
Public-bug tests use inconsistent assertion shapes and are not automatically proven to
fail on the base revision. Release publication also lacks the exact-artifact gate tracked
by #232. A separate manifest and evaluator platform would duplicate the existing pytest
suite without closing either gap.
Expected behavior
#312 standardizes exact public-boundary regression assertions in owner-local pytest tests,
and #313 proves that each new regression fails on base and passes on head. #318 measures
coverage once and exposes one aggregate project CI gate. #319 adds a small bounded set of
properties, #320 makes accuracy data strict and expands labeled cases, and #232 ensures
publication consumes the verified artifact.
Existing tests remain with their owning components. This tracker does not require a case
manifest platform, workspace DSL, changed-surface map, canary corpus, or scheduled
fuzz/mutation system.
Problem
Public-bug tests use inconsistent assertion shapes and are not automatically proven to
fail on the base revision. Release publication also lacks the exact-artifact gate tracked
by #232. A separate manifest and evaluator platform would duplicate the existing pytest
suite without closing either gap.
Expected behavior
#312 standardizes exact public-boundary regression assertions in owner-local pytest tests,
and #313 proves that each new regression fails on base and passes on head. #318 measures
coverage once and exposes one aggregate project CI gate. #319 adds a small bounded set of
properties, #320 makes accuracy data strict and expands labeled cases, and #232 ensures
publication consumes the verified artifact.
Existing tests remain with their owning components. This tracker does not require a case
manifest platform, workspace DSL, changed-surface map, canary corpus, or scheduled
fuzz/mutation system.