Switches the Release workflow trigger from `push: tags` to the GitHub
`release: published` event, and adds a `workflow_dispatch` input for
manual backfills of an existing tag.

release-please creates tags with GITHUB_TOKEN, which by design does not
trigger downstream workflows. That's why v0.9.4 shipped with only
release notes and no binary assets (#424), while v0.9.3 (tag pushed
manually) worked. Reacting to the `release` event published by
release-please fires the workflow reliably for future releases.

The workflow_dispatch path accepts an existing tag and routes it
through checkout, packaging, asset upload, and the Homebrew tap update
so v0.9.4 can be backfilled after this lands.

Fixes #424

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Co-authored-by: haosenwang1018 <haosenwang1018@users.noreply.github.com>

…ance (#427)

Document LLAMA_CPP_PATH and LLAMA_SERVER_PORT environment variables for
llama.cpp integration. Clarify that MLX downloads map to mlx-community
repos. Add note that end users update the model database by upgrading
the package rather than running make update-models.

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Bumps [docker/login-action](https://github.com/docker/login-action) from 3.7.0 to 4.1.0.
- [Release notes](https://github.com/docker/login-action/releases)
- [Commits](docker/login-action@c94ce9f...4907a6d)

---
updated-dependencies:
- dependency-name: docker/login-action
  dependency-version: 4.1.0
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* feat: persist filter state across sessions

Saves all filter selections (fit, availability, TP, sort, installed-first,
search query, and all multi-select popup filters) to ~/.config/llmfit/filters.json
so they are restored on next launch. Follows the existing theme persistence
pattern. Multi-select filters are stored by name rather than index position,
so changes to the model database between runs won't corrupt saved state.

Closes #428

* fix: save filters on quit and remove redundant re_sort on startup

Address PR review feedback:
- Move save_filters() from apply_filters() to quit handler to avoid
  disk I/O on every filter change and ensure all paths persist state
- Replace re_sort() with apply_filters() on startup since all_fits is
  already sorted during construction
- Fix rustfmt formatting

* fix: use dirs::config_dir() for platform-correct config path

Replace manual HOME/USERPROFILE env var lookup with dirs::config_dir()
which handles macOS ~/Library/Application Support, Linux XDG, and
Windows AppData correctly.

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>