**Brief approach**I’d build a static-analysis tool that parses the central JSON token file, indexes token values, and scans web/mobile repositories (JS/TS, Swift, Kotlin, XML, CSS/Sass) to map usages. Then it reports unused tokens, likely duplicates (same value across different names), and produces a migration plan linking token names → code references and suggested consolidations.**Pseudocode**python
# Pseudocode
load tokens = parse_json("tokens.json") # {name: value}
index_value_to_names = invert(tokens) # {value: [names]}
refs = {} # {token_name: [file:line,...]}
for repo in repos:
for file in repo.files:
text = read(file)
for token_name in tokens:
if token_pattern(token_name) in text: # pattern tuned per platform
refs[token_name].append(file:line)
unused = [n for n in tokens if refs[n] is empty]
duplicates = {v: names for v,names in index_value_to_names.items() if len(names)>1}
# Migration suggestions: pick canonical name (e.g., longest-used or design-pref)
for value,names in duplicates:
canonical = choose_canonical(names, refs)
suggest replacing other names -> canonical with refs report
generate_report(unused, duplicates, refs)
