Classify leaf mentions in inference-axis-coverage report

LeafCoverageReport.parseMentions (R104) joins each sealed leaf simple-name against every roadmap *.md body via a \b<simpleName>\b regex. The match is undifferentiated: backticked code spans, code-fenced blocks, and bare prose mentions all collapse into the same Roadmap cell. Two consequences. First, every roadmap edit that names a leaf in any form drifts inference-axis-coverage.adoc and trips the verify-leaf-coverage-report CI gate, which is the regen-friction tax R104 deferred. Second, a reviewer reading the column has no way to sanity-check a match ; Field against FieldType is excluded by \b, but a phrase like "the field type" cannot be told apart from a deliberate Field symbol reference without re-reading the source spec body.

The post-processor already has the full sealed-leaf inventory at parse time, so this is a typing problem, not a data-availability problem. Classify each mention into a sealed Mention.{Backticked, CodeFenced, Prose} (or equivalent) at parse time and surface the kind in the report ; either as a separate column or by collapsing prose-only matches into a paler "weak mention" rendering. That changes the column’s contract from "any token-level overlap" to "this spec references this leaf as a code symbol", which is the intent the column was added for.

Out of scope for this item: changing the regen-friction itself (a workflow change to auto-regen on roadmap touches is a separate question), and any change to the sealed-leaf-inventory parser. This item is purely about the mention join.