docs(skills): drop the usage line; the author stamp is plain provenance

CLAUDE.md

@@ -810,8 +810,7 @@ hand before each release. Work through this list against the sources above:
       `src/agent/memoryNotes.ts`; the placeholders named on `agents/memory` and
       `agents/index` match `src/agent/prompts.ts`
 - [ ] `agents/skills` — the `manage_skills` default (on) matches
-      `builtInToolDefaults.ts`; the usage-line example under "Where a skill
-      came from" matches the wording in `src/skills/usageSummary.ts`
+      `builtInToolDefaults.ts`
 - [ ] `agents/widgets` — header keys match `WidgetSpec` in
       `src/widget/widgetSpec.ts`, `libs` matches `WIDGET_LIBS` in
       `src/widget/widgetLibs.ts` (and the trace types the Plotly build carries),

src/content/docs/agents/skills.md

@@ -142,12 +142,8 @@ queue. Creating a skill is the same action as attaching it.
 A skill the agent creates is stamped `author: agent` under `metadata` in its
 frontmatter, the same key the bundled skills use to name their author. That is
 the whole provenance record, and it travels with the note through sync and
-copies.
-
-The Agent editor's **Custom** section shows each custom skill's origin and use
-on one line, for example *Created by the agent · used 3 times, last 2 days ago
-· revised once*. The counts live in plugin data, not in the note, so loading a
-skill never rewrites the file. Deleting the skill clears them.
+copies. Nothing in the plugin acts on it; it is there so you can tell which
+skills started with the agent.
 
 :::caution
 A skill the agent creates may only request tools from a fixed read-only