Find your next
Claude Code tool

Interface design guidance for Claude Code.

Animated scenes that respond to tool calls.

Live view of subagents, tools and permission waits.

Version-specific library documentation in your workflow.

A pull request review workflow that filters findings using confidence scores.

Git workflow commands for creating commits, pushing branches and opening pull requests.

A guided feature workflow that explores the codebase, designs a solution and reviews the result.

Create local rules that warn about or block matching tool actions.

A toolkit for authoring plugins with skills for structure, hooks, commands, agents and MCP integration.

Specialized review agents for tests, comments, types, error handling and code quality.

Edit warnings and model-assisted reviews for common vulnerability patterns.

Connect TypeScript and JavaScript language intelligence to Claude Code.

Python type checking and code intelligence through the Pyright language server.

Microsoft’s browser tools, packaged as a Claude Code marketplace integration.

GitHub repository tools exposed through a marketplace integration.

A context-window forecast drawn above the Claude Code terminal prompt.

Preview what selected risky shell commands could affect before deciding to proceed.

Step through recorded file-edit calls from the last turn in a review pane.

Author lifecycle hooks with guidance on event selection, configuration and validation.

Guidance on clear skill triggers, concise instructions and progressive disclosure.

Authoring guidance for connecting tool servers to a Claude Code plugin.

A Hookify skill for defining readable rules with event and pattern conditions.

Trace code paths and explain how an existing feature is implemented.

Plan a feature implementation that follows the existing codebase’s structure.

The focused commit command from the Commit Commands package.

Coordinate the PR Review Toolkit’s specialized review lenses from one command.

Track watched GitHub pull requests above the prompt with reviews, required checks and merge state.

A breathing animation appears while Claude works, with optional breathing-phase text in the spinner.

A native session dashboard shows context, usage, tool latency, agents and changed files, with a companion CLI.

Open a Chromium browser beside the transcript for websites, local previews and optional agent browsing.

Animated cartoons above the prompt turn Claude’s work into scenes narrated by additional Sonnet or Haiku requests.

Show account limits and context use above the prompt, with reset forecasts and optional context compaction.

Render GitHub-style alerts, checklists, strikethrough and supported Mermaid diagrams in assistant replies.

Maintain an interactive architecture diagram beside the conversation as planning and implementation progress.

Rewrite em dashes in assistant replies, file writes and edits, and selected Git or GitHub CLI commands.

Switch the main conversation’s model and reasoning effort with keyboard shortcuts and a status footer.

Pin newly spawned non-fork subagents to a source-configured model with a persistent on/off setting.

Estimate cold prompt-cache cost and keep a cache warm through billed background model requests.

Source link only. Billed requests: a detected cold write can automatically arm at least 3 hours of model warming. It sends transcript data to the configured provider. No guaranteed savings.

An idle server-farm game above the prompt reacts to completed turns, test commands and failed tools.

Source link only. License unspecified. This is an attributed source-link listing; no install command or code redistribution is provided.

Create Obsidian notes with wikilinks, properties, embeds and callouts.

Write reusable Playwright scripts for browser checks, screenshots and multi-step flows.

Review PostgreSQL queries, indexes, schemas, connection handling and row-level security.

Review React and Next.js code for request waterfalls, bundle size and rendering work.

Examine code changes for security regressions with Git history and affected callers.

Design meaningful property tests and assess reduced counterexamples.

Ask side questions about the current session in a separate, tool-less pane.

Keep session tasks, open decisions and supported output links beside the transcript.

Select repository profiles for global instructions, Skills, agents and tools.

Guide development through design, implementation planning, testing and review.