CSV Anonymizer v1.0.20

• Add run-scoped pseudonym mapping so repeated source values reuse the same replacement and distinct readable names avoid duplicate assignments while pool capacity remains.
• Expand readable first/last-name pools, preserve first/last/full-name consistency through shared semantic domains, and report avoided candidate collisions plus pool-exhaustion fallbacks.
• Add a Tokenize strategy for opaque tok_... values and upgrade deterministic pseudonyms to keyed HMAC-SHA256 with seed-sensitivity notes in the privacy report.
• Improve the privacy report UI and serialized report fields with unique pseudonym counts, repeated-source reuses, opaque-token counts, avoided collisions, and exhausted pseudonym pools.