Claude Skills by wormeyman

Audit a web page, local file, or published artifact against WCAG 2.2 AA, then fix what it finds. Use whenever the user asks whether something is accessible, says "a11y", "WCAG", "ADA", "Section 508", "screen reader", "color contrast", "keyboard navigation", "tab order", or asks to run Lighthouse for accessibility - and also proactively right after building or restyling any page, because contrast and keyboard defects ship silently. Covers the theme matrix (one Lighthouse run only tests whichev...

Put a decision to the user as an AskUserQuestion prompt with grounded options, never as a prose question. Use when you are about to ask the user anything whose answer changes what you do next - which approach to take, merge or hold, fix now or defer, which target to pick. Use it especially when finishing a long report, because that is where the habit breaks. Covers what counts as a decision, how to write options worth choosing between, and the tool mechanics: option count, headers, multiSelec...

Audit a project's dependencies for available updates, removals, security issues, and major-version features worth adopting. Use this whenever the user asks what can be updated, upgraded, removed, replaced, or is outdated; asks whether a major version would bring useful new features; asks to review the packages, libraries, crates, gems or dependencies in a project; or mentions dependency hygiene, supply-chain risk, abandoned packages, or automated update tooling like Dependabot or Renovate - e...

Use when a page, artifact, report, or static site must go online but stay restricted to named people - someone asks to "password protect" a page, share a document with specific emails only, gate a preview or staging site, put authentication in front of something that has no backend, or keep a confidential deliverable off the public internet. Also use when a published Claude artifact needs an audit trail of who opened it, and whenever a private page is about to be deployed before its access co...

Evidence-driven refactoring assessment and execution for any language. Use this whenever the user wants code cleaned up, simplified, restructured, deduplicated, or made more readable - and also when they say "this file is a mess", "reduce the complexity here", "we have tech debt", "DRY this up", "split this god object", "is this worth refactoring?", or ask for a refactor survey of a repo. Use it BEFORE adding a feature to code the user calls messy, so the restructuring lands as its own review...

Choose an OS-enforced sandbox for running code you do not trust, then prove the sandbox actually holds before running anything in it. Use whenever you are about to execute, build, install, or test something whose contents you have not verified: a contributor's branch or fork PR, a dependency you are auditing, a downloaded binary or archive, a devcontainer or Dockerfile from a repo, a proof-of-concept, or fixture input during a security review. Also use when a workflow tells you it needs "no e...

Generate a self-contained "resume this work" handoff prompt and copy it to the clipboard, so the user can paste it into a fresh session that has zero prior context. Use this whenever the user asks for a prompt to continue/resume/pick up work later, to "pbcopy a prompt", to hand off to a new session or agent, or says things like "give me a prompt to keep going in a new chat", "write me a handoff", "copy a prompt so I can resume this tomorrow", or "I want to start fresh on this" - even if they ...