Third-Party Agent Skills · Browse by Task
Use a focused instruction when the task is to change code: write a test, investigate a failure, improve a React implementation, or refactor without changing behavior.
Coding Skills guide decisions inside an existing codebase. They do not supply a runtime, a repository, or permission to edit production systems. Read each source for its assumptions before applying it.
The source organizes React and Next.js performance rules into eight impact-ranked categories. That makes it useful for a focused review of fetching, bundles, server rendering, client state, and re-renders—not a substitute for profiling a real page.
The source is a prioritized collection of mobile rules. It is particularly helpful where web-oriented React advice misses native list virtualization, gesture performance, safe areas, or image handling.
Guides feature or bug-fix work through tests first and a red–green–refactor loop.
Uses a diagnosis loop for difficult failures and performance regressions before committing to a fix.
Asks for evidence-ranked hypotheses when a failure is ambiguous, intermittent, or performance-related, before any edit.
Structures code extraction, consolidation, or cleanup around verification that existing behavior remains intact.
This is an independent directory. Check the publisher's source for installation, tools, and availability.
Start with TDD when the expected behavior can be stated and checked before implementation.
Use diagnosis or investigation guidance to rank evidence before editing the suspected code.
Choose React-specific guidance for performance issues, or Safe Refactor when behavior must remain unchanged.
Choose the next task, not just the next publisher.
From Instruction to Code
Bring the task, repository context, tests, and review decisions into one workspace.
