fix(plugins): enable aliases using manifest names

[liuhao1024]

What does this PR do?

• Store the manifest-declared plugin name when enabling a plugin by its installed directory name.
• Remove stale directory aliases from enabled/disabled config sets during enable/disable.
• Add a regression test for directory name vs manifest name mismatches.

Related Issue

N/A

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)

Changes Made

• See commit messages for detailed changes

How to Test

1. Run pytest tests/ -q — all tests should pass
2. Verify the specific scenario described above is resolved

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: macOS 26.4.1

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture and workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A

[Tranquil-Flow]

Hi @liuhao1024 — the PR correctly adds _resolve_plugin_config_name() to both cmd_enable and cmd_disable, but the test only covers the enable path (TestCmdEnable.test_enable_by_directory_stores_manifest_name). The disable path is just as likely to be exercised — a user who enabled ping_island by directory name and later wants to disable it would expect _get_disabled_set() == {"ping-island"} rather than {"ping_island"}.

Suggest adding a symmetric test:

class TestCmdDisable:
    def test_disable_by_directory_stores_manifest_name(self, ...):
        cmd_disable("ping_island")
        assert _get_disabled_set() == {"ping-island"}

Mirrors the enable test and locks in that the resolver runs on both sides.

[liuhao1024]

@Tranquil-Flow Good catch — the symmetric disable test was already added in 73943e3d6. It tests the full enable→disable roundtrip with directory alias resolution:

class TestCmdDisable:
    def test_disable_by_directory_stores_manifest_name(self, tmp_path, monkeypatch):
        ...
        cmd_enable("ping_island")
        assert _get_enabled_set() == {"ping-island"}
        cmd_disable("ping_island")
        assert _get_disabled_set() == {"ping-island"}
        assert "ping-island" not in _get_enabled_set()

All 55 plugin tests pass clean. Thanks for the thorough review!

[teknium1]

Thanks for the focused plugin-alias fix. This automated hermes-sweeper review found that current main has since implemented a broader canonical-key solution.

• 2b89afec79f67d4e0aeb50cb3cb16d9853d0150f normalizes manifest names, directory/leaf aliases, and nested plugin keys through hermes_cli/plugins_cmd.py:746.
• cmd_enable and cmd_disable persist the loader's canonical key and reconcile stale aliases at hermes_cli/plugins_cmd.py:814 and hermes_cli/plugins_cmd.py:915.
• Regression coverage includes manifest-name alias cleanup and symmetric enable/disable behavior in tests/hermes_cli/test_plugins_cmd_enable_disable_nested.py:109 and :148.
• The replacement shipped in v2026.6.19.

The PR's requested guarantee is therefore already present on main, with broader nested-plugin handling.