legend-skills

Keep code, tests, documentation, configuration, and other changes as tight, clear, performant, and minimal as possible without weakening behavior. Use during implementation to prevent unnecessary complexity, after behavior is proven to consolidate working changes, before handoff for a full branch audit, and when asked to simplify, tighten, clean up, reduce verbosity, remove complexity, or minimize existing work.

Commit the current agent's changes by default, or all working-tree changes when explicitly requested, with clean logical grouping and repository commit-message conventions. Use when the user asks to commit changes, commit this work, commit all changes, split changes into commits, propose commit messages, or wait for "go" before committing.

Opt-in evidence-first causal diagnosis for bugs, browser or app/device failures, flaky behavior, and performance regressions. Activate only when the user explicitly invokes `$diagnose` or explicitly asks to use the named diagnose skill. Do not activate merely because the user mentions a bug, asks why something failed, requests debugging or a fix, or describes unexpected behavior. Once explicitly invoked, locate likely causes, instrument relevant boundaries with extensive structured logging, reproduce the issue, analyze the collected logs, rank causes with confidence scores, and pursue 100% operational confidence while probes can increase confidence.

Safely start new or continue in-progress Git integration and history operations through verified completion. Use when asked to run or resume a rebase, merge, cherry-pick, or revert; when Git is already in the middle of one of those operations; or for interruption recovery, conflict resolution, ours/theirs interpretation, and deciding when user guidance is required. Continue a detected active operation before considering new work, never choose the integration method, and never guess an unclear next action or resolution.

Opt-in approval-first, whole-working-tree wrapper for the commit skill. Defaults to grouping all changes and requires the commit skill; selected installs must also install commit. Activate only when the user explicitly invokes `$commit-confirm` or explicitly asks to use the named commit-confirm skill. Do not activate merely because the user asks to review or propose commits, ask before committing, or wait for "go".

Build, audit, and fix @legendapp/list and LegendList usage in React, React Native, and web apps. Use for virtualization, scroll blanking, mount cost, data identity, recycling, row invalidation, measurement, chat anchoring, drawDistance, visibility, or Legend List performance regressions.

Build, audit, and fix @legendapp/state usage in React, React Native, and TypeScript. Use for observable ownership, useValue selectors, observer and reactive components, fine-grained subscriptions, effects, persistence and sync, settings or session stores, deprecated use$ or useSelector migration, and reducing re-renders in hot paths.

Implement and review React, React Native, and TypeScript UI for maximum practical render performance through leaf-local updates, stable identities, honest Hook semantics, and minimal effects. Use for components, hooks, callbacks, context, external stores, memoization, list rows, large dependency arrays, callback or state fanout, re-render reduction, and React Compiler-aware performance work.

Iterative diagnose-and-fix workflow that repeatedly uses the diagnose skill to find a proven problem, plan the smallest credible improvement, implement it, re-diagnose, and continue until no useful fix remains or user intervention is required. Requires the diagnose skill; selected installs must also install diagnose. Use when asked to keep debugging, fix and verify, iterate until satisfied, improve after diagnosis, or run a diagnosis/fix/verification loop.