Fallback pattern for reliable file creation when code execution tools fail

Use shell_agent as fallback when execute_code_sandbox fails for Excel file operations

遇到任何 bug、测试失败或异常行为时使用，在提出修复方案之前执行

Turn an idea or objective into a goal package for /goal. Interviews the user, builds a reviewed fact sheet via Plannotator, then explores the codebase to produce an execution plan.

Captures executable contracts and coding conventions into .trellis/spec/ documents. Use when learning something valuable from debugging, implementing, or discussion that should be preserved for future sessions.

High-cost orchestrator workflow for large, high-risk, multi-phase coding efforts with meaningful dependencies and review gates. Do not activate for routine multi-file changes.

Optimize Core Web Vitals (LCP, INP, CLS) for better page experience and search ranking. Use when asked to "improve Core Web Vitals", "fix LCP", "reduce CLS", "optimize INP", "page experience optimization", or "fix layout shifts".

Apply modern web development best practices for security, compatibility, and code quality. Use when asked to "apply best practices", "security audit", "modernize code", "code quality review", or "check for vulnerabilities".

Anti-slop frontend design for web UI — landing pages, portfolios, marketing/product sites, and redesigns. Reads the brief, infers the right direction, and ships interfaces that don't look templated. Real design systems when applicable, audit-first on redesigns, strict pre-flight check. NOT for mobile apps.

当执行 BD/IR 任务需要对外演示材料时创建精美的 HTML 演示文稿——路演 PPT、合作提案、客户 Demo、产品介绍。零依赖单文件输出，可直接通过邮件/微信发送或浏览器演示。

Use for authorized digital forensics including memory dumps, disk timelines, PCAP investigation, artifact triage, and IR evidence preservation.

Reference: fast triage of inbound NDAs into GREEN / YELLOW / RED so the team only spends lawyer time on the ones that need it. Built for sales and BD to self-serve before pinging legal. Loaded by /commercial-legal:review when an NDA is detected.

The Ginkgo mental model for writing Go tests — the one idea that explains everything (Ginkgo builds a spec tree at construction time, then runs it) and its consequences for how you write specs, plus spec independence and the node taxonomy. Use this first when you start working with Ginkgo in a project, or to decide how to approach a Ginkgo task. Routes to the other ginkgo:* skills.

Scan open issues and discussions. Dedupe, prioritize, and propose labels. Updates issue-triage-state.md. L1 propose-only — never auto-label or close.

Takes a structured list of changes from changelog-scan and produces a clean, categorized, user-facing release notes draft. Follows project voice and conventions. Never publishes — only proposes a draft file or section.

Scan recent merges to main (and noteworthy direct commits) since a given window (last tag or date in state). Extract titles, labels, types, linked issues, and signals for breaking changes or security. Produces structured input for a release notes drafter. Use in changelog-drafter loops.

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.

Hugging Face CLI to estimate the required memory to load Safetensors or GGUF model weights for inference from the Hugging Face Hub

Vision-language pre-training framework bridging frozen image encoders and LLMs. Use when you need image captioning, visual question answering, image-text retrieval, or multimodal chat with state-of-the-art zero-shot performance.

Data framework for building LLM applications with RAG. Specializes in document ingestion (300+ connectors), indexing, and querying. Features vector indices, query engines, agents, and multi-modal support. Use for document Q&A, chatbots, knowledge retrieval, or building RAG pipelines. Best for data-centric LLM applications.