fix: reject empty normalized case slugs

[rogerchappel]

Summary

• reject case identifiers whose normalization produces an empty slug
• preserve invalid-name errors through create, load, and add storage APIs
• cover new, add, and show without creating or mutating .clipcase/case
• document accepted identifiers and normalization behavior

Verification

• Tests or checks run: npm run release:check; bash scripts/validate.sh; git diff --check
• Manual review completed: confirmed all three commits use Roger Chappel <miscanalysis@gmail.com> as author and committer

Risk Level

• Low
• Medium
• High

Notes: Valid identifiers retain their existing slug normalization. Punctuation-only identifiers now fail with Invalid case name instead of aliasing the literal case slug.

Commit-by-commit:

• test: cover empty case slugs adds CLI and library regression coverage.
• fix: reject empty normalized case slugs validates before storage reads or writes and preserves validation errors.
• docs: define accepted case identifiers documents the allowed slug characters and normalization rules.

Rollback Plan

• Revert the three commits in reverse order.

Human Decision Needed

• None
• Maintainer review
• Product/design decision
• Security/privacy review
• Other: