fix(*): stop the skill block messages naming a key the write may not use

The same branch made set_skill_blocked reuse whichever spelling the config
file already carries, so the key it writes is now conditional. Two lines
that report what the write just did were not: the log line, and the result
line of raven skill block and raven skill unblock, both named
skillForge.blocklist unconditionally.

On a file spelling the block skill_forge the write lands on
skill_forge.blocklist while both messages say skillForge.blocklist, so an
operator sent to their own config greps for a key it does not carry.

Report the setting rather than the file key. That is what the sibling
set_sentinel_nudge_quota already does: it respects the same two spellings
and logs "sentinel nudge quota patched" with no key path at all. The
--help text and the module docstring keep naming skillForge.blocklist,
because they describe the command rather than the write, and that sibling
names one canonical spelling in its own summary line the same way.

Co-authored-by: Claude (claude-opus-5[1m]) <noreply@anthropic.com>

Author: LivXue

raven/cli/skill_commands.py

@@ -140,7 +140,7 @@ def skill_block(name: str = typer.Argument(..., help="Skill name / slug to refus
     from raven.config.update import set_skill_blocked
 
     blocklist = set_skill_blocked(name, True)
-    console.print(f"[green]Blocked[/green] {name!r}. skillForge.blocklist = {blocklist}")
+    console.print(f"[green]Blocked[/green] {name!r}. Skill blocklist = {blocklist}")
     console.print("[dim]Takes effect on the next agent/gateway start.[/dim]")
 
 
@@ -150,7 +150,7 @@ def skill_unblock(name: str = typer.Argument(..., help="Skill name / slug to all
     from raven.config.update import set_skill_blocked
 
     blocklist = set_skill_blocked(name, False)
-    console.print(f"[green]Unblocked[/green] {name!r}. skillForge.blocklist = {blocklist}")
+    console.print(f"[green]Unblocked[/green] {name!r}. Skill blocklist = {blocklist}")
     console.print("[dim]Takes effect on the next agent/gateway start.[/dim]")
 
 

raven/config/update.py

@@ -245,7 +245,7 @@ def _apply(_text: str | None) -> tuple[str | None, tuple[list[str], bool]]:
     current, wrote = atomic_update(path, _apply)
     if wrote:
         logger.info(
-            "config/update: skillForge.blocklist now {!r} ({} {!r})",
+            "config/update: skill blocklist now {!r} ({} {!r})",
             current,
             "blocked" if blocked else "unblocked",
             name,

tests/test_cli_skill_commands.py

@@ -210,6 +210,26 @@ def test_skill_unblock_removes_from_config(tmp_config: Path) -> None:
     assert _read_blocklist(tmp_config) == ["other"]
 
 
+def test_skill_block_reports_the_block_it_wrote_on_a_snake_config(tmp_config: Path) -> None:
+    """The result line must not name a key this write did not use.
+
+    On a snake_case config the write lands on ``skill_forge.blocklist``, so a
+    line naming ``skillForge.blocklist`` sends the operator to grep for a key
+    their file does not carry.
+    """
+    import json
+
+    tmp_config.write_text(json.dumps({"skill_forge": {}}), encoding="utf-8")
+
+    r = runner.invoke(app, ["skill", "block", "tag-memory"])
+
+    assert r.exit_code == 0
+    assert "tag-memory" in r.stdout
+    assert "skillForge" not in r.stdout
+    data = json.loads(tmp_config.read_text(encoding="utf-8"))
+    assert data["skill_forge"]["blocklist"] == ["tag-memory"]
+
+
 def test_skill_unblock_absent_is_noop(tmp_config: Path) -> None:
     r = runner.invoke(app, ["skill", "unblock", "ghost"])
     assert r.exit_code == 0