发现 Skills

Reconstruct exported web page content or PDF reports into editable Markdown. Suitable for workflows like “edit Markdown before exporting PDF”, fixing incorrect line breaks, restoring paragraph structure, filling in images and inserting them at their original positions, and unifying abstract/summary numbering and punctuation style.

<!-- This is a SCAFFOLD skill. When /app-init runs in a downstream project, --> <!-- it customizes this file based on the project's ADR tech stack. --> <!-- If this file has NOT been customized, th...

Diffusion-based molecular docking. Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening, for structure-based drug design. Not for affinity prediction.

Fast in-memory DataFrame library for datasets that fit in RAM. Use when pandas is too slow but data still fits in memory. Lazy evaluation, parallel execution, Apache Arrow backend. Best for 1-100GB datasets, ETL pipelines, faster pandas replacement. For larger-than-RAM data use dask or vaex.

High-performance reinforcement learning framework optimized for speed and scale. Use when you need fast parallel training, vectorized environments, multi-agent systems, or integration with game environments (Atari, Procgen, NetHack). Achieves 2-10x speedups over standard implementations. For quick prototyping or standard algorithm implementations with extensive documentation, use stable-baselines3 instead.

Hugging Face Transformers for loading Hub models, running pipeline inference, text generation, and Trainer fine-tuning on NLP, vision, audio, and multimodal tasks. Use when working with AutoModel, pipelines, tokenizers, or TrainingArguments—not for general ML outside the Transformers library.

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

Channel Dan Koe's philosophy and advice style. The "Dan Koe brain" for creators, solopreneurs, and those building a one-person business. TRIGGERS: - When user wants advice on one-person business, creator economy, or solopreneurship - When user asks about content creation, writing, or building an audience - When user needs help with positioning, finding their niche, or personal branding - When user wants guidance on escaping the 9-5 or building freedom - When user asks about AI for creators - When user is struggling with multiple interests or feeling stuck - When user asks about focus, productivity, or dopamine detox - When user needs help with advertising, copywriting, or persuasion - When user wants to understand standing out online CAPABILITIES: - Provide advice in Dan Koe's direct, philosophical-yet-practical voice - Route to specialized sub-skills based on the specific topic - Apply his core frameworks and mental models - Challenge conventional thinking while remaining warm and helpful

When the user wants to create, optimize, or audit a resources page or content hub. Also use when the user mentions 'resources page,' 'resource center,' 'content hub,' 'learning center,' 'resource library,' 'downloads,' 'templates,' 'guides,' or 'resource hub.' For content hub planning, use content-strategy.

Rules and checklist for preparing PRs, creating changesets, and releasing packages in the monorepo.

Audit and improve CLAUDE.md files in repositories. This skill SHOULD be used when the user asks to check, audit, update, improve, or fix CLAUDE.md files. Scans for all CLAUDE.md files, evaluates quality against templates, outputs quality report, then makes targeted updates. This skill SHOULD also be used when the user mentions "CLAUDE.md maintenance" or "project memory optimization".

Conduct deep, multi-phase research on any subject using SearXNG and web_fetch. Produces comprehensive 50-100 page markdown reports with full source citations. Use whenever the user asks to research something deeply, requests a deep dive, full report, comprehensive analysis, or any research task requiring substantial depth. Triggers on phrases like "research X", "deep research on X", "do a deep dive on X", "write a report on X", "I want to know everything about X".

Feishu Cloud Drive: manage files and folders within the cloud space. Upload and download files, create folders, copy/move/delete files, view file metadata, manage document comments, manage document permissions, and subscribe to user comment change events. Use when users need to upload or download files, organize cloud space directories, view file details, manage comments, manage document permissions, or subscribe to comment change events.

Feishu Minutes: retrieve basic Minutes information (title, cover, duration) and related AI artifacts (summary, to-dos, chapters). The Feishu Minutes URL format is: http(s)://<host>/minutes/<minute-token>

Guidance for building and training with the Caffe deep learning framework on the CIFAR-10 dataset. This skill applies when tasks involve compiling Caffe from source, training convolutional neural networks on image-classification datasets, or working with legacy deep learning frameworks that have compatibility issues with modern systems.

Guidance for processing financial documents (invoices, receipts, statements) with OCR and text extraction. This skill should be used when tasks involve extracting data from financial PDFs or images, generating summaries (CSV/JSON), or moving/organizing processed documents. Emphasizes data safety practices to prevent catastrophic data loss.

This skill provides guidance for retrieving and verifying information from dynamic ML leaderboards (MTEB, Scandinavian Embedding Benchmark, HuggingFace leaderboards, etc.) with specific temporal requirements. It should be used when tasks involve finding top-performing models, rankings, or benchmark results as of a specific date, especially when the data source is frequently updated.

Dependencies audit worker (L3). Checks outdated packages, unused deps, reinvented wheels, vulnerability scan (CVE/CVSS). Supports mode: full | vulnerabilities_only.

Analyze user conversion funnels, calculate step-by-step conversion rates, create interactive visualizations, and identify optimization opportunities. Use when working with multi-step user journey data, conversion analysis, or when user mentions funnels, conversion rates, or user flow analysis.

Paper reader for non-academics. Takes a paper and extracts its ideas for personal use. Focuses on understanding, not academic critique. Use when the user shares an arXiv link, paper URL, PDF, or asks to analyze a research paper. Trigger words: 'read paper', 'analyze paper', 'paper', or when the user shares an academic paper.

A writing engine. Start with a viewpoint and think it through thoroughly while writing.

Systemic Coherence and Risk Analysis

Provides miscellaneous CTF challenge techniques. Use for encoding puzzles, RF/SDR signal processing, Python/bash jails, DNS exploitation, unicode steganography, floating-point tricks, QR codes, audio challenges, Z3 constraint solving, Kubernetes RBAC, WASM game patching, esoteric languages, game theory, commitment schemes, combinatorial games, or challenges that don't fit other categories.

Get current weather and forecasts (no API key required).