refresh-catalogslisted

# Refresh Catalogs Manually regenerate catalog indexes for memory, shared memory banks, diary, skills, and resources. This skill invokes the catalog dispatcher (`generate_catalogs` via `"${CLAUDE_PLUGIN_ROOT}/scripts/generate_catalog.py"`) to rebuild catalog files used by the context manager for fast routing. ## Usage By default (no arguments), all enabled catalogs are regenerated using state-aware skipping — only catalogs whose source content has changed since the last run are rebuilt. Memory and diary catalogs are always processed (mandatory). Skills and resources catalogs only run if enabled in the plugin configuration. ### Flags - **`--force`** — Force regeneration of all enabled catalogs, bypassing state-aware skipping. Ignores content hash state and regenerates regardless of whether sources have changed. Pass `--force` flag through to the dispatcher. - **`--dry-run`** — Preview mode. Reports what catalogs would be regenerated or skipped without writing any files, modifying `.generation-state.json`, or making LLM calls. Dry-run output shows which generators would run and which would be skipped. No side effects. - **`--only <generators>`** — Selectively regenerate specific catalogs. Provide a comma-separated list of generator names: `memory`, `banks`, `diary`, `skills`, `resources`. Example: `--only memory,diary` runs only those two generators. **`--only` is an explicit override: a generator named here runs even if its `enable_*` config flag is off** (e.g. `--only re