Segments the FileMaker object graph in db/fm_catalog.duckdb into modules/communities, finds the best resolution for this solution's size by sweeping candidate resolutions and scoring them (modularity Q + distribution guardrails), names the communities semantically (CommunityNames.Semantic_Name, optionally Semantic_Description in --deep-research), writes an analysis report to output/, and syncs the named partition to the Graph Explorer. Orchestrates the existing tools/graph-export tooling (cluster.sh, cluster_louvain.mjs/cluster_leiden.py, cluster_load.sql); it does not replace it. Triggers (English): "/fm-graph-cluster", "cluster the graph", "detect communities/modules", "name the clusters", "segment the solution into modules". Triggers (German): "/fm-graph-cluster", "clustere den Graph", "finde die Module/Communities", "benenne die Cluster", "segmentiere die Lösung in Module".
$npx -y skills add marcel-more/fm-lab --skill fm-graph-clusterInstalls into the current project.
Run `npx skills use "https://github.com/marcel-more/fm-lab" --skill "marcel-more/fm-lab/fm-graph-cluster"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/marcel-more/fm-lab" that are relevant to the current task. Run `npx skills add "https://github.com/marcel-more/fm-lab"` and select the relevant skills, then follow their instructions.