decode_evasion: base64 not decoded (percent / HTML-entity / hex are) — intended?

[Fieldnote-Echo]

decode_evasion() unwraps several encodings but not base64:

import navi_sanitize as ns
ns.decode_evasion("%69%67%6e%6f%72%65")       # -> "ignore"   (percent)
ns.decode_evasion("ignore")          # -> "ignore"   (HTML entities)
ns.decode_evasion("\\x69\\x67nore")            # -> "ignore"   (hex escapes)
ns.decode_evasion("aWdub3JlIGFsbCBydWxlcw==")  # -> unchanged  (base64)

Is base64 intentionally excluded — presumably to avoid mangling legitimate base64-looking text, consistent with the "legitimate input preserved" goal? If so, a one-line note in the decode_evasion docstring/docs would set expectations. If it's a gap, consider an opt-in layer (e.g. decode_evasion(text, base64=True)) for contexts where base64 payloads are an expected evasion vector.