发现 Skills

Use when user asks to "create a tour", "add onboarding flow", "implement guided walkthrough", "set up Flowsterix", "add tour steps", "configure step targeting", "handle async content in tour", "use devtools to capture elements", "grab elements for tour", or needs help with tour overlays, popovers, step advance rules, TourHUD configuration, or DevTools element grabber.

Guide development with modern Docker Compose, Docker Bake, and ECR for building and managing container images.

A skill to check region-by-region support for Anthropic Claude models on AWS Bedrock. When a user inputs model names like "sonnet 4.7", "opus 4.7", or "haiku 4.5", it queries the 10 major AWS regions (US, Europe, Asia Pacific) to determine whether the model is available and returns the list of supporting regions along with the actual model IDs. Use cases include: (1) "Which regions can I use sonnet 4.7 in?", (2) "Tell me the supported regions for opus 4.7", (3) "Which regions can I use claude haiku 4.5 in on AWS Bedrock?", (4) requests to check the availability of a specific Claude model.

Google Agent Development Kit (ADK) Python Cheatsheet. Comprehensive reference for building, orchestrating, and deploying AI agents.

Use the gspace CLI and MCP tools for Google Workspace operations (Drive, Gmail, Docs, Sheets, Slides, Calendar, Tasks). Use these tools when working with Google Workspace URLs (docs.google.com, drive.google.com, sheets.google.com, slides.google.com, mail.google.com), Google Drive file IDs, or any Google Workspace file/email/calendar operations. Supports both CLI commands (run via Bash) and 40+ MCP (Model Context Protocol) tools.

Use xsv for fast CSV data processing with selection, filtering, statistics, joining, sorting, and indexing for high-performance data manipulation.

Expert guidance for debugging code, analyzing errors, and systematic problem-solving. Use when troubleshooting bugs, understanding error messages, or investigating unexpected behavior.

Customize the retail agent for different use cases. Use when adapting for different business types, industries, or output formats, or when modifying prompts, adding verticals, or changing the analysis focus.

This skill should be used when the user asks about their saved content, bookmarks, starred articles, highlights, reading history, tweets, or knowledge base in Huntly sqlite database. Triggers include "my favorites", "knowledge base", "what did I save", "my bookmarks", "my library", "what did I save".

Docker 容器化配置指南

Transform essay-to-speech output into complete presentations with multiple output formats. Use when converting talk tracks to slides, generating presentation decks, or creating video-ready content from spoken word material.

Render, build, generate, produce, or export a Domo client proposal as a polished, signature-quality PDF (.pdf) suitable for offline review, enterprise procurement workflows, formal RFP responses, and any moment that calls for a flat, brand-clean artifact a client can read on a plane, forward to legal, or print on letterhead. Use this skill when the browser-print PDF produced by the **Export to PDF** button inside the interactive proposal isn't polished enough — when you need to suppress the browser's URL/timestamp chrome, add a Domo-branded footer with page numbers, and guarantee identical output regardless of which browser the recipient happens to open the file in. This skill is the polished-PDF companion to `domo-interactive-proposal-builder`. Both consume the same source-of-truth engagement content JSON — two outputs from one payload: 1. `domo-interactive-proposal-builder` → interactive HTML proposal (review) 2. `domo-proposal-pdf-renderer` → polished offline PDF (RFP/procurement) Numbers, hour ranges, ROI math, solution shape, deliverable detail — identical across the two formats because they both derive from the same JSON. MANDATORY TRIGGERS: render proposal pdf, generate proposal pdf, polished pdf, bulletproof pdf, signature-quality pdf, rfp pdf, procurement pdf, offline proposal, headless pdf, server-side pdf, playwright pdf, domo-branded pdf, proposal pdf for enterprise, proposal pdf for legal review, build the pdf for, generate the pdf for, render the pdf for. The output is a PDF rendered by headless Chromium against the same `template.html` that the interactive proposal renders to. Every panel is flattened, every solution card is force-expanded, every deliverable's detail blocks (What it includes / Who it's for / Outcome & value) are visible. The Pacific Drift palette is preserved via `print-color-adjust: exact`. No browser-imposed header. A Domo-branded footer with page numbers is rendered by Playwright's footer template (not the page itself, so it stays fixed at the bottom of every page regardless of content flow). Document metadata (Title, Author, Subject) is set so the PDF reads cleanly in document managers and email previews. Do NOT use this skill for: the casual self-service PDF most clients can produce by clicking the Export to PDF button inside the interactive proposal (that's already wired into `domo-interactive-proposal-builder` and is sufficient for ~95% of clients), Discovery Briefs (use `domo-discovery-brief`), or any artifact that isn't a Domo client proposal rendered from the shared engagement content JSON.

Parse documents into Markdown using MinerU Agent lightweight APIs (no token required). Use this skill whenever the user asks to convert PDF, image, DOCX, PPTX, XLS, or XLSX into Markdown quickly, especially in AI agent workflows, local file conversion, batch conversion to a Markdown folder, or recovery from MinerU lightweight errors such as signed upload 403s or 20-page page-limit failures.

SQL query optimization, indexing strategies, and EXPLAIN analysis for Lemline's multi-database architecture (PostgreSQL, MySQL, H2). Use when debugging slow queries, designing indexes, optimizing outbox patterns, or ensuring cross-database compatibility.

Analyzes and optimizes npm package dependencies. Handles outdated packages, npm audit findings, security vulnerabilities, dependency updates, unused dependency removal, and devDependencies recategorization. Recommends the lt-dev:npm-package-maintainer agent via /maintain commands. Activates for "update packages", "npm audit", "check dependencies", "security fix", or package.json optimization. NOT for @lenne.tech/nest-server version updates (use nest-server-updating).

High-performance Rust web crawler with stealth mode, LLM-ready Markdown export, multi-format output, sitemap discovery, and robots.txt support. Optimized for content extraction, site mapping, structure analysis, and LLM/RAG pipelines.

Query JASPAR for transcription factor binding site (TFBS) profiles (PWMs/PFMs). Search by TF name, species, or class; scan DNA sequences for TF binding sites; compare matrices; essential for regulatory genomics, motif analysis, and GWAS regulatory variant interpretation.

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Letta Code's capabilities with specialized knowledge, workflows, or tool integrations.

Create flexible boundaries for projects - not rules or requirements, but bumpers that guide without constraining. Use when starting projects, when feeling overwhelmed by possibilities, or when needing structure without suffocation.

Analyze CI build and test status from Azure DevOps and Helix for dotnet repository PRs. Use when checking CI status, investigating failures, determining if a PR is ready to merge, or given URLs containing dev.azure.com or helix.dot.net. Also use when asked "why is CI red", "test failures", "retry CI", "rerun tests", "is CI green", "build failed", "checks failing", or "flaky tests". DO NOT USE FOR: investigating stale codeflow PRs or dependency update health, tracing whether a commit has flowed from one repo to another, reviewing code changes for correctness or style.

Writes Claude Code status line scripts. Use when creating, customizing, or debugging statusline configurations.

Deploy ML models to edge devices with resource constraints

Develop Godot editor tools for workflow optimization

Configure deadzone calibration for controller sticks