docs(data-modeling): restore the Text-branch secret edge, pointed at secret

Round-1 rework: deleting the Text-branch credential edge outright (previous
commit) left readers who answer "Text? yes" for an api_key/db_password —
the table's own `secret` examples — with no secret-value edge at all on the
branch they actually walk; the only surviving `secret` edge hangs off
`Structured Data?`, which that reader answers "no" to.

Point the Text branch at `secret` instead: `TEXT -->|Sensitive secret
value| S5[secret]`. The Structured-branch edge stays too — an api_key is
text and a credential blob is structured, and a decision tree with two
questions legitimately reaching one answer is ordinary, not ambiguous. `S5`
already has a style rule from the Structured branch, so no new style line
is needed.

Re-ran the full A2 edge-vs-table audit (36 edges now): the new TEXT->secret
edge matches the `secret` table row; zero disagreeing edges.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012GKcPZbMoGq7WPzKLfRBTU

Author: claude

content/docs/data-modeling/field-type-decision-tree.mdx

@@ -33,6 +33,7 @@ flowchart TD
     TEXT -->|Email address| T5[email]
     TEXT -->|Website URL| T6[url]
     TEXT -->|Phone number| T7[phone]
+    TEXT -->|Sensitive secret value| S5[secret]
 
     %% Number branch
     NUM -->|Any number| N1[number]