Design workload creation-failure explanations (#411) #515

Problem

#411 asks for a read-only explanation of why a selected workload's Pods are not being created, from recorded creation failures and the namespace's ResourceQuota and LimitRange objects. Before code, it needs these design decisions:

• which workloads it opens on, and which reads it makes (workload, its ReplicaSets, Events, quotas, LimitRanges);
• what the projections keep, since a Pod template carries literal environment values;
• how a recorded rejection is classified and tied to a quota or LimitRange without overclaiming;
• how quantities are shown and compared without naive string comparison or misleading unit conversion;
• what absence may conclude, and how stale quota status is reported;
• how Explain workload creation failures with ResourceQuota and LimitRange evidence #411 splits into pull requests.

Goal

A design spec under docs/superpowers/specs/ for the maintainer to review before any implementation. Merging it approves the design; each delivery slice then gets its own sub-issue and pull request.

Part of #411.