AI Skills
Installable SKILL.md packages that extend AI coding agents with reliable workflows and domain-specific judgment — ranked by installs, sourced from skills.sh.
tao-train-fast-foundation-stereo
Real-time stereo depth estimation using FastFoundationStereo (FFS), the distilled bp2 commercial variant of
distill-memory
Capture breakthrough moments and valuable insights as searchable memories in your knowledge base.
hyva-alpine-component
Write CSP-compatible Alpine.js components for Hyvä themes in Magento 2. This skill should be used when the user wants to create Alpine components, add interactivity to Hyvä templates, write JavaScript for Hyvä themes, or needs help with Alpine.js patterns that work with Content Security Policy. Trigger phrases include "create alpine component", "add interactivity", "alpine for hyva", "x-data component", "csp compatibility", "csp compliant javascript".
tao-route-visual-changenet-samples
Routes the weakest VCN samples (output of `tao-analyze-gaps-visual-changenet`) into per-augmentation-module
chinese-git-workflow
国内 Git 平台配置参考——Gitee、Coding.net、极狐 GitLab、CNB 的 SSH/HTTPS/凭据/CI 接入差异与镜像同步配置。仅在用户显式 /chinese-git-workflow 时调用，不要根据上下文自动触发。
omero-integration
Microscopy data management platform. Access images via Python, retrieve datasets, analyze pixels, manage ROIs/annotations, batch processing, for high-content screening and microscopy workflows.
tao-train-sparse4d
Sparse4D for multi-camera temporal 3D object detection and tracking. Uses sparse queries with deformable
plaid-fintech
Expert patterns for Plaid API integration including Link token
web-haptics
Add haptic feedback to web apps using the web-haptics library. Use when building mobile-facing UIs that need tactile feedback on buttons, toggles, forms, pickers, and other interactive elements.
java-spring-development
Java Spring Boot development guidelines with best practices for building robust, secure, and maintainable enterprise applications
summarization
Create effective summaries by matching summarization type to purpose, audience, and context. Use when asked to summarize, create TLDR, condense content, or create executive summaries. Keywords: summary, TLDR, condense, executive summary, abstract.
tao-train-deformable-detr
Deformable DETR for 2D object detection. Uses deformable attention for efficient multi-scale feature processing,
glycoengineering
Analyze and engineer protein glycosylation. Scan sequences for N-glycosylation sequons (N-X-S/T), predict O-glycosylation hotspots, and access curated glycoengineering tools (NetOGlyc, GlycoShield, GlycoWorkbench). For glycoprotein engineering, therapeutic antibody optimization, and vaccine design.
tao-mine-aoi-images
Runs the DEFT embed-then-mine workflow for VCN AOI iterations — embeds the gap-analysis target parquet, embeds a source pool, and mines nearest-neighbour source images for downstream augmentation. Use as the immediate next step after `tao-route-visual-changenet-samples` when expanding a real-image augmentation queue from the mining subset.
tao-run-on-lepton
DGX Cloud Lepton managed GPU compute platform with run/status/cancel interface. Use when submitting TAO jobs
mcp-builder
Create MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. The quality of an MCP server is measured by how well it enables LLMs to accomplish real-world tasks.
tao-train-bevfusion
BEVFusion for multi-sensor 3D object detection. Fuses LiDAR point clouds and camera images in bird's-eye-view
tao-train-oneformer
OneFormer for universal image segmentation. Unifies panoptic, instance, and semantic segmentation with a
hubspot-integration
Expert patterns for HubSpot CRM integration including OAuth
tao-train-ocdnet
OCDNet for scene text detection. Detects arbitrary-oriented text regions in natural images using a
tao-validate-dataset-format
Run `tao-daft validate` to check NVIDIA TAO DAFT datasets for structure, schema, and cross-reference errors. Do
qwen-agent
Delegate menial, well-scoped coding tasks to a cheap Qwen-backed subagent via the `claude-9arm` command instead of burning Claude tokens/quota. Use when the work is mechanical and low-risk — bulk renames, formatting, boilerplate, find-replace, grep-style search & summarization, reading/condensing logs or files, test/docstring/comment scaffolding, or running builds/linters/tests and reporting pass-fail. Also use when the user says "use qwen", "delegate this", "send it to 9arm/qwen", or "do this cheaply". Do NOT use for architecture, design, debugging judgment, security-sensitive edits, or anything needing this conversation's context.
cold-email
When the user wants to write, improve, or build a sequence of B2B cold outreach emails to prospects who haven't asked to hear from them. Use when the user mentions 'cold email,' 'cold outreach,' 'prospecting emails,' 'SDR emails,' 'sales emails,' 'first touch email,' 'follow-up sequence,' or 'email prospecting.' Also use when they share an email draft that sounds too sales-y and needs to be humanized. Distinct from email-sequence (lifecycle/nurture to opted-in subscribers) — this is unsolicited outreach to new prospects. NOT for lifecycle emails, newsletters, or drip campaigns (use email-sequence).
wcag-accessibility-audit
Comprehensive web accessibility audit using WCAG 2.1/2.2 guidelines. Evaluate compliance across 4 POUR principles (Perceivable, Operable, Understandable, Robust) with A, AA, AAA conformance levels.
ralph
Convert PRDs to prd.json format for the Ralph autonomous agent system. Use when you have an existing PRD and need to convert it to Ralph's JSON format. Triggers on: convert this prd, turn this into ralph format, create prd.json from this, ralph json.
tao-analyze-gaps-vlm-bcq
Extract false-positive and false-negative gaps from VLM binary-classification-question (BCQ, yes/no) predictions.
finishing-a-development-branch
当实现完成、所有测试通过、需要决定如何集成工作时使用——通过提供合并、PR 或清理等结构化选项来引导开发工作的收尾
research-proposal
>
ssh
SSH remote access patterns and utilities. Connect to servers, manage keys, tunnels, and transfers.
tao-train-rtdetr
RT-DETR (Real-Time DEtection TRansformer) for 2D object detection. Designed for real-time inference with