nM vs nm handling

[iraiosub]

maps "nm" → "nM". So any token matching the pattern number + nm is reported as needing to be nM

[iraiosub]

CASE_AMBIGUOUS_UNITS = {"m", "mm", "nm", "µm", "μm"}

Proposed solution:

Fixer: never touch case for m, mm, nm, µm. Fix spacing and the µ character only.
Validator: when it sees one of those four in lowercase, emit a confirm-this message rather than a preferred form — e.g. "Line 42: 100 nm — nanometre or nanomolar? Use nM if concentration is meant."

Yes, all four files need it — same line in each:

• Use standard units and spacing, for example \37 °C, 10 µL, 1 mM, 100 nM.

It gives 1 mM and 100 nM as examples of "standard units" without saying they're the concentration forms, so an assistant normalising a filter spec could reasonably turn 0.22 µm into 0.22 µM. And the disallowed list has nothing covering unit identity — it forbids inventing values and changing reagent names, not changing what a unit means.

Two additions, in .claude/skills/protocol-migration/SKILL.md, .agents/skills/protocol-migration/SKILL.md, .github/copilot-instructions.md and docs/PROMPT.md:

Under allowed normalisation, replace that line with:

• Use standard units and spacing, for example 37 °C, 10 µL, 1 mM, 100 nM. Normalise spacing and the micro character (μ to µ), but do not change the capitalisation of m, mm, nm, or µm: these are length units and mM, nM, µM are concentrations. Keep whichever the source uses.

Under disallowed changes:

• do not change a unit's capitalisation where this would change the quantity, for example nm to nM or µm to µM