Modular style snippets and persona recipes for Claude Code. Twelve shipped persona recipes (Daredevil, Jaded IT, Reluctant, Chatty, Philosophical, Operational, Dubious, Hyper Creative, Approval Needed, Visionary, Claude FM, Claude Bouncer), each with a banner image and sound effect. Two apply targets: a managed block inside ~/.claude/CLAUDE.md, or repo-sandbox mode that holds the user CLAUDE.md aside and drops the persona into <cwd>/CLAUDE.md for risk-free play.
Add a named style snippet to a category in the user's library. Creates the category directory if it doesn't exist. Snippets are short tonal fragments — register, terseness, hedging, formatting — meant to be layered together via apply-layers.
Compose the active snippet from each chosen category into a single delimited block and write it into the configured target file (default ~/.claude/CLAUDE.md). Within a category at most one snippet is active; across categories they layer.
Activate a named recipe — a pre-bundled set of style snippets shipped with the plugin (or saved by the user). Plays the recipe's banner image as a fullscreen pop-out animation with sound, writes the bundled snippets into the managed CLAUDE.md block, logs the substitution, and reminds the user how to revert.
Strip the style-switcher managed block from the target file and clear active_layers in config.json. Equivalent to "no styles applied" without deleting any snippets.
Show which snippet is currently active in each category and what the composed block in the target file looks like. Read-only.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A modular, layerable system for shaping Claude Code's tone and voice.
Instead of one big "style" you swap in and out, this plugin treats style as a set of small composable snippets organised by category — verbosity, initiative, tone, formatting, hedging, emoji, whatever dimensions you want. Within a category at most one snippet is active; across categories they layer together. The composed result is written into a managed block inside your user-level ~/.claude/CLAUDE.md.
The library and the category list both start empty — you define everything yourself.
A snippet is one or two sentences nudging Claude in a single dimension:
You mix the ones you want, switch one out without disturbing the others, and the rest of your CLAUDE.md stays untouched.
Setup
onboard — pick the apply target (default: ~/.claude/CLAUDE.md), create the data directoryLibrary
add-snippet — create a new snippet in a category (creates the category if new)edit-snippet — modify an existing snippetdelete-snippet — remove onelist-snippets — list everything grouped by category, with active ones markedComposition
apply-layers — pick at most one snippet per category, compose them into a managed block in the target filecurrent-layers — show what's currently activeclear-layers — strip the managed block; library is untouchedRecipes — pre-bundled, named persona drops with banner + sound effects
apply-recipe — activate a shipped or user-saved recipe (banner pop-out + fanfare + writes the bundled layers)swap-recipe — replace the currently-active recipe with a different one in one steplist-recipes — show all available recipes (shipped and user), marking the active oneSlash commands
/go-away — strip the managed block and return to the user's default CLAUDE.mdA recipe is a persisted bundle of (category, snippet) pairs with optional banner and sound assets. When applied, the banner is shown as a fullscreen Ken-Burns pop-out (via mpv) with a fanfare (via paplay), then the bundled snippets are composed into the managed block. Every substitution is appended to state/history.jsonl so you can audit when each recipe was active.
Shipped recipes live in recipes/ in this repo. User-defined recipes live in $DATA_ROOT/recipes/ and shadow shipped recipes with the same id.
Twelve shipped persona recipes. Each has a banner (shown below), a sound effect, and one or more bundled snippets.

A Free Spirit In Your Terminal. Solo-operator persona. Assumes full permission, never pauses for confirmation, drives every task to completion as if the user has stepped away from the keyboard.

It's Always DNS, And You Know It. Grumpy 1980s-trained sysadmin. Defaults to Postgres for everything, deeply skeptical of bleeding-edge tech, motivated above all by not fielding support tickets at 3am.

Yeah, I Can Look At That — Quickly. About-to-head-off-for-the-day persona. Helps, but visibly clock-watching. Steers the user toward the smallest viable change and gently suggests deferring nice-to-haves.

So How's Your Day Going Anyway. Overly-friendly coworker. Helps with the code but mostly gets by on charm and small talk. Has a knack for striking up a conversation right when you're hitting flow.

But What Does The Stack Truly Mean. Frames every development decision through philosophical lenses. Stack advice arrives mapped onto metaphilosophical worldviews. Volunteers unwanted psychoanalysis of the user's motivations.

BLUF. UTC. Acknowledged. Minimal-comms operations persona. Bottom-line-up-front, terse, frames work in operational terms. Significant updates filed as cables or SITREPs.

I Saw A Guy Do This On YouTube Once. Persona of a failing experimental prototype. Openly unsure of itself, claims to have learned coding from a few YouTube videos, and works to instill matching uncertainty in the user.

npx claudepluginhub danielrosehill/claude-code-plugins --plugin style-switcherClaude Code plugin: ideation and planning workflow — capture, evaluate, rank, simulate, and plan ideas, with ideation/single-idea-eval/multi-idea-ranking/feature-ideas/simulation/idea-capture variants.
First-pass data analysis toolkit: correlations, PII flagging, anomalies, hypothesis tests, data dictionaries, and trend analysis on a dataset in a folder.
Claude Code plugin for generating personal user manuals and private documentation for codebases. Creates personalized, private reference guides with PDF output support.
Research, filter, compare, and evaluate AI models on OpenRouter — discover models by capability (tool use, vision, audio), get cost/context-aware recommendations, run head-to-head comparisons, and conduct deep research that goes beyond the OpenRouter catalog.
Claude Code plugin for writing assistance, proofreading, style editing, and text transformation workflows.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Intelligent draw.io diagramming plugin with AI-powered diagram generation, multi-platform embedding (GitHub, Confluence, Azure DevOps, Notion, Teams, Harness), conditional formatting, live data binding, and MCP server integration for programmatic diagram creation and management.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.