refactor(reload-skills): queue note for next turn, drop cache invalidation + agent tool

Salvage-follow-up to @shannonsands's /reload-skills PR. Trims the feature to
match the design: user-initiated rescan, no prompt-cache reset, no new
schema surface, no phantom user turn, and the next-turn note carries each
added/removed skill's 60-char description (not just its name).

Changes vs the original PR:

* Drop the in-process skills prompt-cache clear in reload_skills(). Skills
  are invoked at runtime via /skill-name, skills_list, or skill_view —
  they don't need to live in the system prompt for the model to use them.
  Keeping the cache intact preserves prefix caching across the reload so
  /reload-skills pays no cache-reset cost. (MCP has to break the cache
  because tool schemas must be known at conversation start; skills do not.)

* Drop the skills_reload agent tool and SKILLS_RELOAD_SCHEMA from
  tools/skills_tool.py, plus the four skills_reload enumerations in
  toolsets.py. No new schema surface — agents can already see a freshly-
  installed skill via skill_view / skills_list the moment it's on disk.

* Replace the phantom 'role: user' turn injection with a one-shot queued
  note. CLI uses self._pending_skills_reload_note (same pattern as
  _pending_model_switch_note, prepended to the next API call and cleared).
  Gateway uses self._pending_skills_reload_notes[session_key]. The note
  is prepended to the NEXT real user message in this session, so message
  alternation stays intact and nothing out-of-band is persisted to the
  transcript.

* reload_skills() now returns added/removed as
  [{'name': str, 'description': str}, ...] (description truncated to 60
  chars — matches the curator / gateway adapter budget). The injected
  next-turn note formats each entry as 'name — description' so the model
  can actually reason about which new skills to call without running
  skills_list first.

* Only emit the note when the diff is non-empty. On empty diff, print
  'No new skills detected' and do nothing else.

* Tests rewritten to cover the queue semantics, the description payload,
  and a regression guard that the prompt-cache snapshot is preserved.

tests/cli/test_cli_reload_skills.py

@@ -1,6 +1,14 @@
-"""Tests for the ``/reload-skills`` CLI slash command (``HermesCLI._reload_skills``)."""
+"""Tests for the ``/reload-skills`` CLI slash command (``HermesCLI._reload_skills``).
 
-from unittest.mock import MagicMock, patch
+The CLI handler prints the diff (name + description) for the user and —
+when any skills were added or removed — queues a one-shot note on
+``self._pending_skills_reload_note``. The note is prepended to the NEXT
+user message (see cli.py ~L8770, same pattern as
+``_pending_model_switch_note``) and cleared after use, so no phantom user
+turn is persisted to ``conversation_history``.
+"""
+
+from unittest.mock import patch
 
 
 def _make_cli():
@@ -15,13 +23,18 @@ def _make_cli():
 
 
 class TestReloadSkillsCLI:
-    def test_reports_added_and_removed(self, capsys):
+    def test_reports_added_and_removed_and_queues_note(self, capsys):
         cli = _make_cli()
         with patch(
             "agent.skill_commands.reload_skills",
             return_value={
-                "added": ["alpha", "beta"],
-                "removed": ["gamma"],
+                "added": [
+                    {"name": "alpha", "description": "Run alpha to do xyz"},
+                    {"name": "beta", "description": "Run beta to do abc"},
+                ],
+                "removed": [
+                    {"name": "gamma", "description": "Old removed skill"},
+                ],
                 "unchanged": ["delta"],
                 "total": 3,
                 "commands": 3,
@@ -30,19 +43,28 @@ class TestReloadSkillsCLI:
             cli._reload_skills()
 
         out = capsys.readouterr().out
-        assert "Added: alpha, beta" in out
-        assert "Removed: gamma" in out
+        assert "Added Skills:" in out
+        assert "- alpha: Run alpha to do xyz" in out
+        assert "- beta: Run beta to do abc" in out
+        assert "Removed Skills:" in out
+        assert "- gamma: Old removed skill" in out
         assert "3 skill(s) available" in out
-        # An informational message should be appended to the conversation
-        # so the model picks up the diff on the next turn.
-        assert len(cli.conversation_history) == 1
-        msg = cli.conversation_history[0]
-        assert msg["role"] == "user"
-        assert "Skills have been reloaded" in msg["content"]
-        assert "alpha" in msg["content"]
-        assert "gamma" in msg["content"]
 
-    def test_reports_no_changes(self, capsys):
+        # Must NOT pollute conversation_history — alternation-safe.
+        assert cli.conversation_history == []
+
+        # One-shot note queued with system-prompt-style formatting.
+        note = getattr(cli, "_pending_skills_reload_note", None)
+        assert note is not None
+        assert note.startswith("[USER INITIATED SKILLS RELOAD:")
+        assert note.endswith("Use skills_list to see the updated catalog.]")
+        assert "Added Skills:" in note
+        assert "    - alpha: Run alpha to do xyz" in note
+        assert "    - beta: Run beta to do abc" in note
+        assert "Removed Skills:" in note
+        assert "    - gamma: Old removed skill" in note
+
+    def test_reports_no_changes_and_queues_nothing(self, capsys):
         cli = _make_cli()
         with patch(
             "agent.skill_commands.reload_skills",
@@ -57,10 +79,10 @@ class TestReloadSkillsCLI:
             cli._reload_skills()
 
         out = capsys.readouterr().out
-        assert "No changes detected" in out
+        assert "No new skills detected" in out
         assert "1 skill(s) available" in out
-        # Nothing changed — don't pollute history.
         assert cli.conversation_history == []
+        assert getattr(cli, "_pending_skills_reload_note", None) is None
 
     def test_handles_reload_failure_gracefully(self, capsys):
         cli = _make_cli()
@@ -73,5 +95,5 @@ class TestReloadSkillsCLI:
         out = capsys.readouterr().out
         assert "Skills reload failed" in out
         assert "boom" in out
-        # Failure must not append a misleading "skills reloaded" note.
         assert cli.conversation_history == []
+        assert getattr(cli, "_pending_skills_reload_note", None) is None