phyr97

Deep research across web, codebase, and knowledge domains with auto-scaling. Use when: "research", "deep research", "investigate", "compare", "analyze across", "what are best practices for", "how does X compare to Y", "survey options for". Supports web research, codebase analysis, knowledge synthesis, and mixed mode.

Compare the current Claude Code version against the last verified plugin baseline and propose modernizations. Use when: "modernize plugin", "check cc update", "neue cc version", "plugin updaten", "claude code changelog gegen plugin". Manual gate — never auto-commits or pushes. Approval via MODERNIZATION-PROPOSAL.md.

Run a structured expert panel to turn an open question into a defensible recommendation with explicit trade-offs. Use when user faces an open decision (X vs Y), brainstorming need, or wants multi-perspective analysis.

Multi-mode code review with agent-based analysis. Use when: "review", "code review", "/review", "check this diff", "review my changes", "review this branch", "review this commit". Supports diff review, staged review, commit review, file review, and plan/text review.

Analyze how well Claude Code plugins perform across sessions. Use when: "plugin insights", "plugin performance", "how are my plugins doing", "plugin analysis", "plugin report", "which plugins work well". Generates an HTML report from JSONL session transcripts.