Strengthen closure criteria and cross-round memory in iterative review loops

[ZenusZhang]

Summary

The analyzed loop showed that the process was effective at finding real issues, but slower at closing them. Repeated repairs of the same category of gap suggest the methodology needs stronger closure discipline.

Observed Patterns

• Iterations often produced real progress, but several rounds revisited the same class of missing condition.
• Reviewer feedback was useful and usually correct, but it more often identified gaps than stated a clear closure check.
• Summaries became less decision-oriented over time, which made handoff less crisp.

Suggested Improvements

1. Define the minimum closed state at the start of each round, not after partial progress is already made.
2. Require each review comment to include an explicit pass/fail closure condition.
3. Keep a lightweight cross-round ledger of unresolved issues so repeats are easier to spot.
4. Make round summaries focus on three items: done, blocked, next priority.
5. Force reviewer attention to blocking items before narrower hardening work.
6. Add a stabilization limit so long runs of only boundary tightening trigger a methodology reset.

Expected Benefit

These changes should reduce repeated repair cycles, improve closure confidence, and make each round easier to use as a handoff artifact.