Rebuilds your entire ADR index and relationship graph in one shot by calling a Node script instead of making hundreds of individual MCP calls. It scans for ADRs in docs folders, handles both YAML frontmatter and markdown heading formats, extracts supersedes/amends/depends-on relationships, and stores everything in the adr-patterns and adr-edges namespaces. You'd reach for this after importing ADRs from another repo or when your graph has drifted from the filesystem. The false-positive filtering is smart enough to ignore issue numbers and commit hashes that look like ADR IDs. Pair it with adr-verify to catch cycles before they become a problem.
npx skills add https://github.com/ruvnet/ruflo --skill adr-index