`lstrip("PMCpmc")` in pubmed_search.py strips characters, not a prefix

[Bug-Keeper]

Pretty sure this is a .lstrip() semantics confusion. In skills/sn-search-academic/scripts/pubmed_search.py:119:

pmc_id = id_elem.text.lstrip("PMCpmc").strip() or id_elem.text

The comment one line up says "去掉 'PMC' 前缀，只保留数字" — strip the PMC prefix and keep only digits. But str.lstrip(chars) doesn't strip a prefix; it strips any leading run of characters contained in chars. So if PubMed ever returned an ID with mixed casing or any leading char that happens to be in "PMCpmc", the result would be wrong:

• "PMC123" → "123" ✓ (by coincidence)
• "PMCC456" → "456" ✓ (because C is in the set)
• "pmpc789" → "789" — also stripped, but that's not a real prefix
• A hypothetical "PMP500" → "500" — bug surface in disguise

In practice today PMC IDs are always PMC + digits, so the bug doesn't fire — but removeprefix expresses intent and is robust:

text = id_elem.text or ""
pmc_id = text.removeprefix("PMC").removeprefix("pmc").strip() or text

Severity: Low (latent — current PMC IDs are well-formed) but worth fixing because the wrong abstraction here is the kind of thing that bites two years later when the API adds a new ID variant.