dotfiles

Query code semantics (symbols, hover, definitions, references, diagnostics) across any language via the lsmcp MCP server, which bridges through kakehashi to real language servers (lua_ls, basedpyright, rust_analyzer, r_language_server, nixd, …). Use when exploring or analyzing a codebase, finding/understanding symbols, or when the first project-wide search misses a language. Prefer these tools over Grep/Read for code analysis.

Run the full multi-stage review pipeline on the current branch — /review (multi-perspective subagents) → codex MCP → PR → Copilot/Gemini bots — fixing findings at each stage until convergence. Use when the user asks to "revise", run the review pipeline, or prepare a branch/PR for merge review.

Run a strict multi-perspective review with fact-checked findings. Use when the user asks for a review, code review, design review, risk review, or wants independent reviewer perspectives before merging or shipping.

Perform deep multi-perspective analysis with problem-specific analytical lenses, cross-validation, and a final recommendation. Use when the user asks to ultrathink, deeply analyze a complex problem, compare strategic options, expose risks, or make a high-impact technical decision.

Stage meaningful diffs and create Conventional Commits with WHY-focused messages. Use whenever making git commits, grouping changes, choosing commit types, or separating structural and behavioral work.

Guide structural code improvements using Kent Beck's Tidy First methodology. Use when seeing messy code, before making behavioral changes, after completing features, documenting deferred cleanup, or discussing when to clean up code.

Guide Test-Driven Development using Kent Beck's Red-Green-Refactor cycle. Use when writing tests, implementing features via TDD, entering RED/GREEN/REFACTOR phase, or following plan.md test instructions.

Guide design decisions using the State > Coupling > Complexity > Code priority framework. Use when evaluating tradeoffs, reviewing design choices, deciding between implementations, or when code volume seems to be prioritized over deeper concerns.

Guide code refactoring using Martin Fowler's catalog of behavior-preserving transformations. Use during TDD REFACTOR phase, when code smells are detected (duplication, long methods, feature envy), when discussing structural improvements, or before behavioral changes (Tidy First).

Manage Architecture Decision Records and ADR-aligned development. Use when user says "we should use", "let's go with", asks "why did we choose X?", implements a feature that may have an ADR, updates ADR status, or needs to check alignment with existing decisions.

Guide logging practices based on Dave Cheney's minimalist philosophy. Use when adding log.Info/Debug/Error/Warn/Fatal calls, reviewing logging code, handling errors with log+return pattern, discussing log levels, or designing error handling strategies.