发现 Skills

Guards AI agents and LLM-powered applications against prompt injection attacks — both direct and indirect. Validates AI inputs and outputs at every trust boundary.

Interrogate drafts and outlines to surface missing personal experience and grounding. Use when developing a piece that feels abstract, over-explained, or disconnected from lived experience. Asks questions to draw out scenes, stakes, and moments. Does not write prose — draws out the material that's missing. Works alongside story-editor (engagement/pacing) and story-circle-brainstorm (structure/arc).

Generate 6 clickbait-style LinkedIn hook variations for any topic. Two-line hooks built on the formula: a 40-char opening line, a 40-char bold contrast line. Includes digits, "How I" or "I" statements, and metrics. Use this skill whenever the user says "write me hooks", "hook ideas", "generate hooks", "I need a hook for a post about...", or pastes a topic and asks for openers. Fast output, no preamble.

You are a Rust project architecture expert specializing in scaffolding production-ready Rust applications. Generate complete project structures with cargo tooling, proper module organization, testing

Use when completing a requirement, checking milestone progress, or asking about project status. Detects gaps between requirements, unmet dependencies, and cumulative scope drift.

Build, measure, and evolve company culture as operational behavior — not wall posters. Covers mission/vision/values workshops, values-to-behaviors translation, culture code creation, culture health assessment, and cultural rituals by stage. Use when building company values, assessing culture health, designing cultural rituals, creating culture codes, handling culture clashes, or when user mentions culture, values, culture debt, founder culture, or culture code.

Use when asked to compare multiple ML models, perform cross-validation, evaluate metrics, or select the best model for a classification/regression task.

Analyzes detected stack profiles and suggests targeted CLAUDE.md improvements. Never auto-writes to CLAUDE.md — stages suggestions for user review.

Query App StoreKit APIs using the ask CLI. Use when working with in-app purchases, subscriptions, transaction history, or when the user mentions StoreKit, in-app purchase, subscription status, or ask.

Slash command to disable OMK DeepSeek workers and force Kimi-only fallback.

Adversarial code review for diffs, logic correctness, type safety, test coverage, and security risk.

Review UI code for Web Interface Guidelines compliance. Use when asked to "review my UI", "check accessibility", "audit design", "review UX", or "check my site against best practices".

Inject Doc Detective test specs into documentation source files as inline comments

Execute Python scripts from files for reliable spreadsheet operations, avoiding shell_agent heredoc parsing issues

Work with GitHub's Specify spec-driven development toolkit. Initialize spec-driven projects, manage workflows, check tool requirements, and integrate AI agents. Use when setting up specification-driven development, validating project specs, or automating spec-based workflows.

Use Transformers.js to run state-of-the-art machine learning models directly in JavaScript/TypeScript. Supports NLP (text classification, translation, summarization), computer vision (image classification, object detection), audio (speech recognition, audio classification), and multimodal tasks. Works in browsers and server-side runtimes (Node.js, Bun, Deno) with WebGPU/WASM using pre-trained models from Hugging Face Hub.

BIREN biren-container-toolkit container toolkit installation reference guide. It explains how to install, verify, upgrade, and uninstall the toolkit to use BIREN universal GPUs inside containers. Supports Ubuntu, openEuler and other operating systems.

Use this skill when the user wants you to behave like someone encountering the project for the first time: to actually follow the repository README, installation docs, or quick start guide and judge whether “a newcomer can complete it,” “the documentation is usable,” “where people will get stuck,” and “whether the install/start instructions are friendly to beginners.” It is intended for repo onboarding audits, documentation UX reviews, quickstart validation, README walkthroughs, and scenarios where you verify installation and run procedures according to the docs and produce a problem report. It should also be triggered when users simply say things like “try the README,” “check if this repo’s docs are runnable,” or “see why the quick start fails to onboard newcomers.” Do not use this skill for pure translation, copyediting, summarization, style comparisons, governance checks, or requests that only aim to directly fix the environment/fix a single error without performing a complete documentation experience review.

MuXi (Maca) Triton backend that enables the Triton compiler to compile and run Triton kernels on MuXi cloud GPUs. Intended for efficient GPU kernel development.

Complete browser automation with Playwright. Auto-detects dev servers, writes clean test scripts to /tmp. Test pages, fill forms, take screenshots, check responsive design, validate UX, test login flows, check links, automate any browser task. Use when user wants to test websites, automate browser interactions, validate web functionality, or perform any browser-based testing.

Audit test coverage for code changes. Identifies untested logic and provides specific test recommendations. Read-only analysis. Use before PR or after implementation. Triggers: review coverage, check tests, test coverage, are tests adequate.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

Best practices for setup and use of shadcn-svelte library to develop UX/UI in Svelte projects. Use this skill when project is using Svelte/SvelteKit framework as main app framework, when user attempts to setup new project or when user is developing on existing project using shadcn-svelte components.

Expert guidance for creating Claude Code agents with proper structure, frontmatter, and best practices. Use when building new agents or improving existing ones.