jordanl17

A/B testing and content experimentation methodology for data-driven content optimization. Use when implementing experiments, analyzing results, or building experimentation infrastructure.

General principles for structured content modeling that apply across CMSs, with Sanity-specific guidance. Use when designing content schemas, planning content architecture, or evaluating content reuse strategies.

Comprehensive Sanity development best practices covering integration guides (Next.js, Nuxt, Astro, Remix, SvelteKit, Hydrogen), GROQ performance, schema design, Visual Editing, images, Portable Text, page builders, Studio configuration, TypeGen, localization, and migrations. Use this skill when building, reviewing, or optimizing Sanity applications.

SEO and AEO (Answer Engine Optimization) best practices including EEAT principles, structured data, and technical SEO. Use when implementing metadata, sitemaps, structured data, or optimizing content for search engines and AI assistants.

<<SKILL_DESCRIPTION>>

A minimal fixture skill for @visill/build tests.

A minimal fixture skill for @visill/build tests.

Render Linear issues as an interactive inline editor for review and edit before saving. Works for fresh drafts AND substantive edits to existing issues (title and/or description). A freeform prompt field captures next-turn instructions. Use when the user wants to draft new Linear issues OR rewrite the title or description of an existing issue inline, including phrases like "draft a Linear issue", "queue up Linear tickets so I can edit", "look at SAPP-1234 and let's edit the description", or "rewrite the description of X". The widget IS the editing surface; do not draft the content as Markdown prose first. Do not trigger for read-only queries (list, search, assign, comment). Do not trigger for single-field swaps ("change the title to X"); apply those inline. Do not trigger for one-shot create-and-save with no review step.

Render Linear issues as an interactive inline editor for review and edit before saving. Works for fresh drafts AND substantive edits to existing issues (title and/or description). A freeform prompt field captures next-turn instructions. Use when the user wants to draft new Linear issues OR rewrite the title or description of an existing issue inline, including phrases like "draft a Linear issue", "queue up Linear tickets so I can edit", "look at SAPP-1234 and let's edit the description", or "rewrite the description of X". The widget IS the editing surface; do not draft the content as Markdown prose first. Do not trigger for read-only queries (list, search, assign, comment). Do not trigger for single-field swaps ("change the title to X"); apply those inline. Do not trigger for one-shot create-and-save with no review step.

Interactive decision tree widget when the user wants help walking a bounded, branching decision. Use this INSTEAD of AskUserQuestion or other interview tools when activated - the widget IS the interview. Triggers: "help me plan...", "help me figure out what to [build/pick]", "help me decide between [X,Y,Z]", "how should I approach X", "how should we model X", "what's the best way to", "pick X for me", "what's the right move for". Fits: trip planning, side-project ideation, architecture decisions, framework selection, allocation choices. The user usually supplies context (scale, budget). If not, ask one clarifying question first - never render cold. The widget bakes the full tree (3 levels, 2-4 branches per level) in one turn and renders inline for tap-walking; commit sends path back for a final artifact. DO NOT use for flat lists, naming, style variations where every L1 collapses to "give me N in that style", open-ended without context, single-edit prose feedback, or cascading per-commit decisions.

Renders multi-section drafts as an interactive widget where each paragraph and bullet is tappable for inline per-unit guidance. Use when the user signals intent to iterate on a draft, either at generation time ("draft an RFC and let me give feedback on each section," "write X so I can iterate on each part") or retroactively on prior multi-section assistant output ("let me give feedback on each part," "make that targetable," "I want to mark up what you just wrote"). Do not trigger for one-shot drafting requests with no stated iteration intent, single-edit feedback like "fix bullet 3," content that is mostly code or tables, or output under roughly three paragraphs.