awesome-claude-code

Runs automated experiment loops: Optuna parameter sweeps or LLM-proposed changes.

1 skill

Generates HTML pages for collecting human ground-truth judgments.

1 skill

Reference for using claude -p as a programmatic LLM backend from Python.

1 skill

Generates app icons and favicons from HTML/CSS via headless Chrome.

1 skill

Maps what a user knows about a topic, then writes a personalized explainer.

1 skill

Guides Playwright end-to-end tests toward semantic locators and auto-retrying assertions.

1 skill

Generates a pedagogical HTML walkthrough of a PR or feature branch.

1 skill

Writes READMEs that explain architecture and design decisions rather than restating types.

1 skill

Reviews React code for unnecessary or incorrect useEffect usage.

1 skill

Generates self-contained HTML diagrams, tables and visual explanations (by nicobailon).

1 skill

Reads all changed files on the current branch to restore context after /clear.

Read-only subagent that explains how code works, with file:line references.

Subagent that finds where code lives (superseded by the built-in Explore agent).

Subagent for structured code review with severity tiers.

Groups the session's changes into logical commits with clear messages.

Lints each file right after Claude edits it and feeds errors back to Claude.

Runs lint and typecheck on staged files and blocks Claude's git commit if they fail.

Blocks Claude Code from committing directly to main or master.

Rebases a stack of dependent PRs onto a target branch.

Researches a codebase question with parallel subagents and writes a research document. Pairs with the codebase-locator and codebase-analyzer plugins.

Self-reviews the current branch before a PR (superseded by built-in /code-review).

Evaluates PR review comments for validity before fixing them.

Detects drift between documentation and code.