All skills

Audits whether a brand's website is technically readable and citation-ready for AI answer engines, then outputs a scored GEO/AEO action plan with 15 buyer prompts.

A meta-skill that walks through designing, writing, validating, and packaging a Claude Code skill.

A Playwright automation skill that sends one prompt to Yuanbao, LongCat, Doubao, Qwen, Gemini, Grok, or MiniMax through a single wrapper command.

Teaches Claude to read and update 15Five performance-management data (users, check-ins, objectives) through the Membrane CLI.

A skill that uses the Membrane CLI to query and automate 1C-Bitrix tasks, users and departments.

A rigid, step-by-step release orchestrator for bundle-plugins: pre-flight audits, doc sync, version bump, CHANGELOG, tag, and GitHub Release.

Auto-checks forms, schemas, auth flows and APIs for personal-data risks before Claude writes the code, mapped to US privacy regulations.

A disciplined refactoring playbook that drives trace-mcp tools to assess risk, map impact, rename atomically, and verify results.

An end-to-end ASO skill covering keyword research, store metadata rewriting, review analysis, ASO scoring and A/B test planning for iOS and Android.

A skill that designs serverless-first AWS architectures, drafts IaC templates, and reviews cost and security.

Convenes a panel of Socrates, Lao Tzu, Watts, Nietzsche, Occam and Feynman to break creative blocks and rethink content strategy.

A six-strategist panel (Schumpeter, Munger, Sun Tzu, Machiavelli, Taleb, Kahneman) researches the market and delivers a structured verdict on business decisions.

Reconstructs where your work stands by mining intent, decisions and rejected approaches from git commit history.

Audits the real, live Google Tag Manager runtime on a site before any event schema work begins.

Compresses natural-language memory files (CLAUDE.md, todos, preferences) into terse caveman-speak to cut input tokens.

Control and query Xiaomi/Mi Home smart-home devices from chat via the mijiaAPI CLI.

Search and manage Claude's persistent memory kept in one portable .mv2 file.

Pushes Claude to build production-grade frontend UIs with a bold, intentional aesthetic instead of generic AI-looking output.

Reads a liarjs fingerprint report and attributes each failing check to the component that actually produced the signal.

Gate your CI build on browser-fingerprint regressions using liarjs baselines and score floors.