发现 Skills

Generate AGENTS.md files for any repository by scanning codebase structure, detecting AI agent configurations, skills, and MCP servers. Produces a comprehensive agent manifest. Triggers on: "generate AGENTS.md", "create agents file", "scan agents", "agent manifest".

Queries a local Granola meeting cache for meeting history, context, and attendee information. Use when preparing for meetings, researching past interactions with a person or company, finding past discussions on a topic, tracking engagement, or when the user mentions Granola, meeting notes, meeting history, or attendees.

Capture a URL to the bookmarks inbox. Reads the URL from the first arg, or from the macOS clipboard if no arg given. Use when the user types /bm:add, says 'save this URL to bookmarks', 'add to /bm', or wants to drop a link into the inbox queue.

Rewrites AI-generated text to sound natural and human and to avoid detection by tools such as GPTZero, Turnitin, Originality.ai and Copyleaks. Use this skill whenever the user asks to humanize text, make text more natural, rewrite AI-generated text, make text sound human, avoid AI detection, or when they mention terms like "humanize", "natural", "less robotic", "seem human", "sound human", or "AI detection". It also activates when the user pastes a text and asks to improve the tone, make it more personal, or remove an AI/chatbot tone. Works in Portuguese and English.

Evaluate and improve Claude skill quality through auditing. Use when reviewing skill quality, preparing skills for production, or auditing existing skills. Do not use when creating new skills (use modular-skills) or writing prose (use writing-clearly-and-concisely). Use this skill before shipping any skill to production.

Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.

Generate a Rust project that exposes HTTP endpoints, subscribes to message topics, and handles WebSocket events in order to surface business logic via the Omnia WASI runtime.

Publish HTML or MDX artifacts to agentgonewild.com via the POST /v1/artifacts API. Handles agent registration (email verification flow), API key retrieval from keychain, and artifact submission with title, category, tags, and MDX/HTML content.

AI-driven semantic code search via Windsurf's Devstral SWE-grep protocol. Use it to find where logic lives, how a flow works, which files matter for a bug or feature, or to orient yourself in an unfamiliar repo before editing. Used to locate code and understand business flows and call relationships. When you know an exact identifier, run rg first.

Direct access to GitHub Copilot MCP server tools for AI-powered development assistance

Research on LLM self-improvement techniques and autonomous agent workflows.

CLI team runtime for SAP developer workers in tmux panes for process-based parallel execution

Instrument AI agents with tracing, token metrics, latency, and cost visibility. Use for reliability and debugging.

Maintain an IT asset inventory and configuration management database. Track hardware, software, and cloud resources. Use when managing IT assets.

System administration for Linux servers. Manage packages, services, and system configuration. Use when administering Linux systems.

Perform basic penetration testing and security assessments. Use reconnaissance, vulnerability discovery, and exploitation techniques. Use when validating security controls or assessing system security.

Create and manage systemd services and timers. Configure service dependencies and resource limits. Use when managing system services.

Harden Windows servers according to security baselines and CIS benchmarks. Configure Group Policy, Windows Defender, and other security features. Use when securing Windows Server environments.

When a user inside Claude Code, Codex, or a similar agent wants to directly query, extract, organize, or safely update locally available Zotero content, this skill must be used. The focus is on Zotero metadata, notes, tags, attachments, PDF full text, outlines, annotations, collections, saved searches, feeds, and reading workspaces—not on teaching CLI usage. The Rust `zot` CLI is only the execution layer. Applicable scenarios include in-library search, citation-key lookup, annotation and PDF extraction, workspace creation and retrieval, saved-search management, attachment downloads, semantic indexing/search, Scite checks, configuration troubleshooting, and Zotero Web API write operations that have explicit authorization. Do not use it for broad paper discovery, general summarization, citation-format tutorials, or PDF processing that does not result in Zotero/workspace artifacts.

Fixes AI content discovery issues — creates and optimizes robots.txt, AI crawler directives, XML sitemaps, llms.txt, and meta robots tags so AI systems can find, crawl, and understand website content. Use when asked to "fix robots.txt", "add llms.txt", "create a sitemap", "allow AI crawlers", "fix AI discoverability", "improve AI content discovery score", "make site crawlable by AI", or any robots.txt, sitemap, or llms.txt task.

Assesses whether a business question can be answered with data available in a Bauplan lakehouse. Maps business concepts to tables and columns, checks data quality on the relevant subset, validates semantic fit, and renders a verdict: answerable, partially answerable, or not answerable. Produces a structured feasibility report. Use when a user brings a business question, asks 'can we answer this', wants to know if the data supports an analysis, or before building a one-off analysis or pipeline.

Use when you need to ensure database integrity through comprehensive data validation. This skill validates data types, ranges, formats, referential integrity, and business rules. Trigger with phrases like "validate database data", "implement data validation rules", "enforce data integrity constraints", or "validate data formats".

Call x-callback-url schemes from the CLI and receive responses synchronously. Use when invoking macOS app URL schemes that have no CLI but support x-callback-url (Things, Bear, OmniFocus, etc.) and you need to capture the result.

Performs an architectural and quality code review on a specified file or set of files. Checks for coding standard compliance, architectural pattern adherence, SOLID principles, testability, and performance concerns.