Claude Skills Collection

A collection of skills for Claude Code and Codex that extend agent capabilities with specialized knowledge and workflows.

Two tightly-scoped packs also live at cdd.dev/skills. Same install convention (npx skills add ckorhonen/<repo>):

swe-skills — 15 swe: skills for engineering judgment work: PR risk review, repo introspection, performance and security audits, incident follow-up, ownership maps. Invoked in conversation.
hone-skills — 8 hone: skills that run on a cadence (daily, weekly, per-PR) to fight code entropy: method brevity, naming clarity, duplication, broken-windows detection.

This repo is the broad collection — wider in scope, more varied in purpose.

About

Here are some skills I've found useful for various projects I've been building. Would love feedback if anyone else finds them useful—have at it! If you're interested in some of my writings on AI and software development, follow me at CDD.dev.

Available Skills

Skills are located in the skills/ directory.

Development & Testing

`agent-browser`

Headless browser automation using agent-browser CLI with element refs optimized for AI agents

Scrape a product catalog with accessibility trees instead of CSS selectors
Automate a multi-step checkout form submission
Run parallel browser sessions to test race conditions in a web app

`agent-engineering`

Battle-tested engineering principles for AI coding agents — plan-first workflow, subagent delegation, self-improvement loops, and verification gates

Set up a plan-first coding workflow that prevents agents from writing code before the design is approved
Create a self-improvement loop that benchmarks agent performance and iterates on prompts
Build a verification gate that runs tests before any commit is accepted

`babysit-pr`

Create, monitor, and shepherd a GitHub pull request through CI, review, merge, and best-effort deployment validation

Open a PR that respects the repo's PR template and required reviewer assignments
Watch CI after each push and summarize failures with fix suggestions
Handle review comments from Greptile and re-request review after fixes are pushed

`bird-fast`

Interact with X/Twitter from the terminal using bird CLI with browser session authentication

Post a thread announcing a new open-source release from a script
Search for mentions of a product and summarize sentiment
Schedule daily tweets from a content queue

`ceo-review`

Review documents and proposals through a CEO lens — technical depth, data-driven decisions, and long-term strategic vision

Review a product spec for whether it aligns with the 3-year platform vision
Critique an engineering proposal on ROI and competitive positioning
Evaluate a go-to-market plan for defensibility and market timing

`ck-review`

Review documents and proposals through the lens of a senior principal engineer — clarity, pragmatism, and shipping velocity

Review an architecture decision record for over-engineering and hidden complexity
Critique a technical spec for ambiguous requirements that will block implementation
Evaluate a proposed API design for consistency with existing conventions

`codex-advisor`

Get second opinions from OpenAI Codex CLI for plan reviews, code reviews, and hard problems

Get a second opinion on an implementation plan before starting a large refactor
Validate security assumptions in a cryptographic module
Unblock a tricky bug you've been stuck on for 30+ minutes

`coreml-optimizer`

Optimize CoreML models for iOS and macOS deployment with quantization, palettization, pruning, and Neural Engine targeting

Optimize a CoreML model for iPhone 15 Pro Neural Engine with 4-bit palettization
Profile inference latency with Instruments and identify Neural Engine fallbacks
Debug accuracy degradation after applying 8-bit quantization to a vision model

`cto-review`

Review through a CTO lens — platform thinking, mobile-first craft, and iterative execution

Review a microservices architecture proposal for platform leverage and team scalability
Critique a mobile app design for offline-first patterns and battery impact
Evaluate a technical roadmap for sequencing risk and iterative milestones

`github-actions-templates`

Create production-ready GitHub Actions workflows for automated testing, building, and deploying

Create a CI/CD pipeline for a monorepo with Nx that only re-tests affected packages
Set up a matrix build for a cross-platform Electron app on macOS, Windows, and Linux
Build a deploy workflow with blue/green environment promotion and rollback

`ios-app-tester`

Test iOS apps using AXe CLI for accessibility auditing, UI automation, and simulator control

Automate a complete onboarding flow test in the iOS Simulator using accessibility identifiers
Audit an app for VoiceOver label gaps and generate a remediation report
Record a test execution video for a critical user journey to attach to a bug report

`macos-apps`

Build professional native macOS apps in Swift with SwiftUI and AppKit, CLI-only without Xcode

Build a native macOS menu bar app with SwiftUI and a persistent popover
Integrate an AppKit NSTableView into a SwiftUI view for complex data display
Set up a Swift Package Manager project with unit tests and a release build pipeline

`mcp-tester`

Test and evaluate MCP server tools in the current session — auditing, validation, and test case generation

Audit all tools exposed by an MCP server for schema correctness and edge case handling
Generate a comprehensive test suite for a new MCP server before shipping it
Validate that an MCP server returns structured errors rather than stack traces for bad inputs

`systematic-debugging`

Systematic approach to debugging bugs, test failures, and unexpected behavior before proposing fixes

Reproduce a flaky test failure deterministically before touching any code
Isolate whether a performance regression is in the database layer or application code
Trace a production 500 error back to its root cause using logs and minimal reproduction

Visual & Design

`ascii-pixel-art`

Transform images into animated ASCII art with subject detection, blurred backgrounds, and cinematic effects (pulse, shine, flicker, hover)

Convert a portrait photo into an animated ASCII art piece with phosphor glow
Generate a retro cyberpunk visualization of a city skyline with scanline effects
Create an interactive web art piece with hover-triggered ASCII animation

`evangelion-design`

Evangelion-inspired NERV-like command-center interfaces with sharp geometry, tactical typography, and danger-first palettes

Build a real-time monitoring dashboard styled after NERV's command center with warning-red accents
Design a terminal UI with sharp hexagonal geometry and alert-state color progressions
Create a login screen with tactical HUD typography and animated scan-line overlays

`gemini-image-generator`

Generate images using Google Gemini API for text-to-image, editing, and multi-image reference

Generate a set of consistent app icons in a flat design style from a text brief
Create marketing banner variants for A/B testing with different color schemes
Produce game sprite sheets for a 2D platformer based on character descriptions

`gemini-visual`

Front-end development assistant powered by Google Gemini for UI analysis, design comparison, and screenshot-to-code

Convert a Figma screenshot into production-ready HTML/CSS with accurate spacing and typography
Compare before/after design mockups and identify all layout regressions
Extract a full color palette and typography scale from a brand screenshot

`imagegen`

Generate and edit images via the OpenAI Image API with a bundled CLI

Generate product photography variants on white backgrounds for an e-commerce listing
Inpaint a person out of a stock photo and replace with a branded background
Batch-generate 10 social media card variants from a template description

`nano-banana`

Generate and edit AI images using Google's Gemini 3 Pro Image model via MCP

Create an illustration for a blog post from a detailed text description
Edit a product photo to change the background to an outdoor setting
Generate a logo concept from a brand brief with multiple style variations

`practical-typography`

Professional typography guidance based on Matthew Butterick's Practical Typography

Audit a marketing landing page for line-length, leading, and font-pairing violations
Choose a type system for a long-form reading app that maximizes legibility
Fix a PDF template that uses body text for headings and has inconsistent spacing

`shadertoy`

GLSL fragment shaders and procedural graphics for Shadertoy and WebGL

Write a raymarched SDF scene with ambient occlusion and soft shadows in GLSL
Implement a procedural noise-based animated background for a website hero section
Port a Shadertoy effect to a Three.js ShaderMaterial for a WebGL project

`skimmable`

Enforce code readability and state minimisation before PRs — reduce arguments, remove optionality, and make diffs shorter

Refactor a function with 8 parameters into a well-named options object before opening a PR
Remove optional state flags that have been defaulting to the same value for 6 months
Shrink a 400-line diff by extracting pure helper functions into a separate module

`tui-designer`

Design and implement retro/cyberpunk/hacker-style terminal UIs with CRT effects, neon glow, and scanlines

Build a terminal-aesthetic dashboard in React with Tuimorphic including CRT scanline and phosphor glow effects
Implement a hacker-style progress screen in SwiftUI using Metal shaders for real-time noise
Design a cyberpunk color palette and typography system for a CLI tool

`ui-design`

Opinionated constraints for building better interfaces — Tailwind, motion/react, Radix/Base UI, and accessibility

Build an accessible modal dialog with Radix UI and Tailwind that traps focus and announces to screen readers
Add enter/exit animations to a sidebar nav using Framer Motion without layout jank
Audit a form for WCAG 2.1 AA compliance and generate a fix list

Marketing & Communications

`app-marketing-copy`

App Store/Google Play listings, ASO keywords, taglines, screenshot captions, ad copy, landing page copy, and email campaigns

Write a 4000-character App Store description for a productivity app with keyword-optimized headers
Generate 10 screenshot caption variants for A/B testing the onboarding flow screens
Craft a launch email sequence (3 emails) for a new iOS app release

`content-trend-researcher`

Analyze trends across Google Analytics, Trends, Substack, Medium, Reddit, LinkedIn, X, and more for data-driven content strategy

Research rising search trends in the "developer tools" niche to plan a quarter's content calendar
Analyze competitor Substack newsletters for topics that drive high engagement
Generate article outlines ranked by search intent and estimated traffic potential

`direct-mail-strategist`

Expert direct mail marketing strategist for copywriting, design, and measurement

Write a 6×9 postcard mailer for a SaaS product targeting SMB owners with a 30-day trial offer
Design a measurement plan with holdout groups for a 50,000-piece mailing
Critique a draft letter for response rate optimization using proven direct mail frameworks

`freshdesk-api`

Freshdesk helpdesk API for building integrations, managing tickets, contacts, and automating support workflows

Build a webhook integration that auto-creates Freshdesk tickets from Stripe payment failure events
Write a script to bulk-close all tickets older than 90 days with a resolved status
Create a dashboard that pulls open ticket counts by product area via the API

`poke-assistant`

Send messages and notifications to Poke via webhook API

Notify yourself when a 2-hour model training job completes on a remote server
Send an alert when a production health check fails with a link to the incident dashboard
Deliver a daily build summary with pass/fail counts and a link to the CI report

`poplar-direct-mail`

Design and send programmatic direct mail using Poplar's HTML templates and API

Build a triggered postcard campaign that mails to users 3 days after they abandon a shopping cart
Create a personalized HTML letter template with dynamic first-name and offer fields
Integrate Poplar's API into a CRM workflow to send win-back mailers to churned customers

`seo-optimizer`

Analyze HTML/CSS websites for SEO optimization and implement best practices

Audit a marketing site for missing meta descriptions, duplicate title tags, and broken canonical URLs
Add structured data (JSON-LD) for a recipe blog to enable rich results in Google Search
Generate a sitemap.xml and robots.txt for a Next.js static site

AI & LLM

`autopredict`

Polymarket prediction market trading agent framework — scan markets, evaluate edges, backtest strategies, and tune parameters

Scan all open Polymarket markets and rank by estimated edge vs. current odds
Backtest a news-momentum strategy against 3 months of resolved markets
Tune confidence threshold parameters to maximize Sharpe ratio on historical trades

`hyperagent`

Self-referential self-improving agent loop — meta-agent modifies task-agent code to optimize any measurable target (based on Facebook Research Hyperagents paper)

Set up a hyperagent loop that iteratively improves a code generation agent's pass@1 score on HumanEval
Run a self-improving prompt optimizer that rewrites its own system prompt based on eval feedback
Build a meta-agent that tunes a classification pipeline to minimize false positives on a labeled dataset

`llm-advisor`

Consult other LLMs (GPT-5.2, GPT-5, Gemini 3) for second opinions on complex problems

Get GPT-5 and Gemini's independent assessments of a distributed systems design before committing to it
Compare how three frontier models approach a tricky regex edge case
Use a cross-model vote to decide between two competing database schema designs

`llm-evaluation`

Implement comprehensive evaluation strategies for LLM applications using automated metrics, human feedback, and benchmarking

Build an automated eval harness for a RAG pipeline using RAGAS metrics (faithfulness, relevance, context recall)
Set up a human preference labeling workflow for comparing two prompt variants at scale
Create a regression benchmark that alerts when a model update degrades performance on golden test cases

`mcp-builder`

Guide for creating high-quality MCP servers in Python (FastMCP) or Node/TypeScript

Build a FastMCP server that wraps the Stripe API with typed tool schemas and proper error handling
Create a TypeScript MCP server that exposes a PostgreSQL database as queryable tools
Design a multi-tool MCP server for a SaaS product's internal APIs with auth middleware

`paperclip`

Interact with Paperclip control plane API for task management, agent coordination, and goal tracking

Create and assign tasks to a fleet of specialized agents via the Paperclip API
Query goal completion status across multiple running agent sessions
Coordinate a multi-agent workflow where downstream agents are triggered by upstream task completions

`prompt-factory`

Generate production-ready mega-prompts across 15 professional domains with quality validation

Generate a system prompt for a customer support agent with escalation rules, tone guidelines, and refusal policies
Create a code review prompt tuned for a TypeScript/React codebase with team-specific conventions
Produce XML-formatted prompts optimized for Claude with chain-of-thought and output schema

Security

`blockchain-auditor`

Security audit smart contracts for exploitable vulnerabilities from an unprivileged context

Audit a Solidity DEX contract for reentrancy vulnerabilities and flash loan attack surfaces
Analyze an ERC-20 token contract's bytecode for hidden mint/burn backdoors
Generate a PoC exploit for an integer overflow in an AMM pricing function and test it on a local fork

`security-best-practices`

Language and framework-specific security reviews for Python, JavaScript/TypeScript, and Go

Review a Python FastAPI service for SQL injection, SSRF, and insecure deserialization vulnerabilities
Audit a Node.js Express app for prototype pollution and path traversal issues
Generate a secure-by-default configuration checklist for a new Go microservice

`security-threat-model`

Repository-grounded threat modeling with trust boundaries, assets, attacker capabilities, and mitigations

Generate a STRIDE threat model for a multi-tenant SaaS API with enumerated abuse paths and mitigations
Document trust boundaries for a mobile app that handles PII and payment data
Write a threat model document for an OAuth 2.0 integration covering token theft and PKCE bypass scenarios

Infrastructure

`cloudflare-manager`

Comprehensive Cloudflare account management for Workers, KV, R2, Pages, DNS, and Routes

Deploy a Cloudflare Worker with KV storage for edge-side session caching
Configure DNS records and route rules for a multi-origin application behind Cloudflare
Set up an R2 bucket with a Worker proxy for private asset serving with signed URL support

`google`

Google Workspace assistant via gog CLI — Gmail, Calendar, Drive, Sheets, Tasks, and Contacts

Draft and send a Gmail message with an attachment pulled from Google Drive
Create a recurring Google Calendar event for a weekly team standup with video link
Update a Google Sheet with the latest data from a CSV export and share it with a team

`markdown-fetch`

Fetch web content as clean Markdown via markdown.new — 80% fewer tokens than raw HTML

Fetch the Stripe API documentation for webhook events as clean Markdown for analysis
Convert a competitor's pricing page to Markdown to compare tier structures
Retrieve a GitHub README as Markdown to summarize a library's features

`subway-info`

Real-time NYC subway information via subwayinfo.nyc REST API

Check real-time arrival times for the A train at West 4th Street
Get current service alerts for the L line before commuting
Find the nearest station to a given address and show live departures

Graphics & 3D

`gsplat-optimizer`

Optimize 3D Gaussian Splat scenes for real-time rendering on iOS, macOS, and visionOS using Metal

Analyze a 2M-splat .ply file and generate a pruning plan targeting 60fps on iPhone 15
Design a LOD scheme for a visionOS experience that switches between detail levels based on distance
Benchmark a Gaussian splat scene with Instruments and identify Metal shader bottlenecks

Media & Video

`video-editor`

Expert guidance for video editing with ffmpeg, encoding best practices, and quality optimization

Transcode a 4K ProRes file to H.265 with optimal CRF settings for streaming delivery
Hardsub an SRT subtitle file onto an MP4 while preserving HDR metadata
Batch-convert a folder of MKV files to MP4 with AAC audio for web compatibility

Web3 & Blockchain

`opensea-api`

Interact with OpenSea's NFT marketplace API for metadata, listings, and events

Fetch all NFTs owned by a wallet address with their current floor price estimates
Track real-time sales events for a collection to alert on significant trades
Pull collection stats (volume, floor, holders) for a competitive market analysis report

Game Development

`playdate-dev`

Build Playdate games in Lua with the Playdate SDK, including game loop, sprites, graphics, input (crank, buttons, accelerometer), and simulator workflow

Build a Playdate puzzle game where the crank controls a rotating mechanism to solve each level
Implement a sprite-based platformer with parallax scrolling on the 1-bit display
Set up a pdxinfo manifest and Simulator build pipeline for submitting to the Catalog

Research & Analysis

`autoresearch`

Run rigorous autonomous experiment loops with explicit hypotheses, repeated trials, structured logs, and local HTML reports

Run 20 iterations of a prompt variant test with structured win/loss logging and a final HTML report
Benchmark two embedding models on a retrieval task with statistical significance testing
Iterate on a hyperparameter sweep for a classification pipeline with keep/discard decisions per run

`qmd`

Local hybrid search for markdown notes and docs

Search a local Obsidian vault for notes related to "authentication patterns" using semantic + keyword hybrid search
Find all notes that mention a specific project name across a large document collection
Retrieve the most relevant design decisions from a markdown wiki for a given technical question

`scientific-critical-thinking`

Evaluate research rigor including methodology, experimental design, statistical validity, and biases

Apply GRADE criteria to assess the evidence quality of a clinical study before sharing its findings
Identify p-hacking and HARKing in a published machine learning paper
Critique the experimental design of an A/B test report for confounds and underpowering

Self-Improvement & Optimization

`continuous-learning`

Auto-monitors interactions to identify learning opportunities and create new skills when patterns repeat

Detect recurring database migration patterns across sessions and auto-generate a migration skill
Surface a gap in agent capabilities after three failed attempts at the same task type
Propose and scaffold a new skill file when a workflow has been manually repeated 3+ times

Planning & Process

`idea-md`

Create or expand Idea.md / IDEA.md files from rough descriptions, repo context, conversation history, or notes

Turn a brainstorm or chat transcript into a durable concept brief before writing a PRD
Expand a rough idea draft into a structured document with problem, users, MVP, and validation steps
Create a repo-grounded Idea.md that reflects existing packages, constraints, and technical leverage

`brainstorming`

Structured exploration of user intent, requirements, and design before implementation

Explore 5 alternative architectures for a feature before picking one to implement
Scope a vague "add notifications" request into specific user stories and acceptance criteria
Map out edge cases and failure modes for a payment flow before writing a line of code

`executing-plans`

Execute written implementation plans in separate sessions with review checkpoints

Run a 12-step database migration plan with a checkpoint after each destructive operation
Execute a feature rollout plan across three services with rollback gates between each deployment
Implement a refactoring plan that touches 20 files with progress tracking and diff review at each checkpoint

`writing-plans`

Create implementation plans from specs or requirements before touching code

Turn a 2-page product spec into a phased implementation plan with clear task dependencies and time estimates
Create an architecture plan for a new microservice that sequences schema, API, and client work
Decompose a "migrate from REST to GraphQL" initiative into safe, reviewable milestones

Meta & Skill Management

`skill-finder`

Proactively discover, install, and create skills from the claude-plugins.dev registry

Install the right skill for an unfamiliar domain on-demand using a "matrix kung-fu" style request
Discover and compare three candidate skills for a task before selecting the best fit
Create a custom skill when no existing registry match covers a recurring workflow

Skill Origins & Attribution

Some skills in this collection originate from external projects:

Design skills (adapt, animate, audit, bolder, clarify, colorize, critique, delight, distill, extract, frontend-design, harden, normalize, onboard, optimize, polish, quieter, teach-impeccable, ui-design) — From Impeccable Design. Systematic, opinionated UI/UX design constraints for building better interfaces with agents. Install the full set from their site.
App Store Connect skills (asc-*) — From app-store-connect-cli-skills by Rudrank Riyam. Skills for managing the full iOS/macOS release pipeline via the asc CLI.

Installation

skills.sh (recommended)

Install any skill with skills.sh:

# Interactive — browse and select skills to install
npx skills add ckorhonen/claude-skills

# Install a specific skill directly
npx skills add ckorhonen/claude-skills --skill <skill-name>

Manual

Copy the skill folder to the appropriate location:

Claude Code

# Project-level (share with team via version control)
cp -r skills/<skill-name> /path/to/project/.claude/skills/

# User-level (available across all projects)
cp -r skills/<skill-name> ~/.claude/skills/

Codex

cp -r skills/<skill-name> ~/.codex/skills/

Contributing

To add a new skill:

Create a new directory in skills/ with your skill name (use kebab-case)
Add a SKILL.md file with YAML frontmatter:

---
name: your-skill-name
description: Brief description of what the skill does
---

# Skill Title

Your skill documentation here...

Submit a pull request

License

MIT License - see LICENSE for details.