Find dead code in hlstatsnext.com — unused exports, orphan/unreferenced files, dead dependencies, unreachable branches, and stale commented-out blocks across the monorepo. Produces a confidence-tiered findings report. Does NOT delete code.
Orchestrate the hlstatsnext.com codegen pipeline (Prisma client + Pothos types + Apollo client types) and detect schema drift. Use after schema changes, after pulling main, or when types appear to disagree with the schema.
Audit the hlstatsnext.com daemon for memory leaks, stability, and fault tolerance — RabbitMQ consumer hygiene, RCON scheduler cleanup, server-state collection growth, shutdown ordering, graceful degradation. Produces a structured findings report. Does NOT fix code.
Generate or update documentation for hlstatsnext.com — package READMEs, ARCHITECTURE.md, JSDoc on exported surfaces, CHANGELOG entries. Use when code needs docs or existing docs are outdated.
Execute an implementation plan step by step for hlstatsnext.com. Use when a plan markdown file exists (in docs/plans/) and the user wants it built, or when directly asked to implement something against an existing plan.
Audit the hlstatsnext.com GraphQL surface (Pothos schema + Yoga server + Apollo client) for security, correctness, performance, and exclusion-list hygiene. Produces a structured findings report. Does NOT fix code.
Analyze feature feasibility against the hlstatsnext.com monorepo and produce a structured implementation plan. Use when the user wants to plan a new feature, schema change, daemon module, or API surface before building it.
Scan the hlstatsnext.com codebase for god classes, mixed concerns, duplication, and readability problems. Produces a prioritized refactoring report. Does NOT write code.