Search AI add-ons

Build interactive HTML artifacts with multiple frontend components.

Choose it for richer application structure; simple single-file artifacts are outside its intended focus.

Apply a coherent color and font theme to an existing artifact.

A theme changes presentation; it does not validate the artifact content or data.

Prepare small animated assets suitable for sharing in Slack.

Output is an animated image; it does not post to Slack or configure a workspace.

Develop and evaluate a reusable skill for a specific workflow.

Meaningful evaluation requires representative tasks and success criteria supplied by the operator.

Create or revise presentation files, including slide structure and notes.

Targets PowerPoint files and templates; other deliverable formats use separate workflows.

Support extraction, assembly and editing tasks for PDF documents.

Scanned inputs can require OCR; inspect the skill dependencies and validate extracted results.

Guide the design of MCP tools for connecting an agent to a service.

A working integration still requires API access, implementation and verification of tool behavior.

Adapt internal updates to a consistent organizational format.

Requires the appropriate communication template and organization-specific context.

Guide visual direction and typography for a new or revised interface.

Focuses on visual design decisions; it does not independently verify application functionality.

Prepare and revise Word documents with deliberate document structure.

Targets Word files; PDF and spreadsheet workflows need their own skills and dependencies.

Structure an iterative process for writing documentation with a collaborator.

Choose it when context and reader feedback are available; it is not a one-shot file converter.

Add focused questions that help a reader inspect an actionable answer.

Designed for selected advice and draft artifacts; routine code and trivial lookups are excluded.

Look up Claude API and SDK details while developing an integration.

This guidance targets Anthropic integration work; using an API can introduce separate usage charges.

Guide the composition of static posters and visual artwork.

Intended for still visual deliverables rather than interactive application behavior.

Apply Anthropic visual identity to documents and other artifacts.

Its brand choices belong to Anthropic; adapt a separate skill for a different organization.

Prepare interactive generative graphics with repeatable random seeds.

Choose this for code-driven art; it requires a p5.js workflow rather than a general image editor.

Find relevant Claude learning resources for an onboarding question.

Useful for learning a feature; it is outside the intended scope of an already active task.