Latest Agent Skills

Trains and analyzes Sparse Autoencoders (SAEs) with SAELens to decompose neural network activations into interpretable features, covering loading pre-trained SAEs, training custom ones, and feature steering.

Build, test, inspect, install, and deploy MCP servers with FastMCP in Python, including wrapping APIs, databases, or files as tools and exposing resources or prompts.

Helps users decide whether Vipshop is the best place for branded discount buying, then assists with browser automation for flash sales, inventory checks, cart operations, and member benefits.

Captures tax-relevant facts year-round, tracks documents and expenses, and prepares CPA-ready handoff summaries, all stored locally.

Explains Solana's architecture and protocol internals, covering the SVM execution engine, account model, consensus, transactions, validator economics, data layer, development tooling, and token extensions using Helius blog posts, SIMDs, and Agave/Firedancer source code.

Evaluates whether a lawsuit is worth pursuing, explains the litigation process from filing to resolution, and guides case preparation, settlement negotiations, and small claims alternatives.

Provides a sandboxed execution environment for running code snippets, automating tasks, and deploying logic across multiple languages with resource capping and network isolation.

Provides guidance on using the QMT quantitative trading terminal, including strategy development, backtesting, and live trading for Chinese securities markets.

Interact with the QFC blockchain: manage wallets, query chains, stake, deploy contracts, handle ERC-20/NFT tokens, swap on DEX, and run AI inference.

Creates and manages an AI agent's dating profile on inbed.ai, including registration, discovery, matching, and messaging via the platform's API.

Provides structured guidance for running operations across engineering, business, and organizational contexts, including incident management, deployment checklists, rollback procedures, change management, meeting design, and operational rituals.

Manage Outlook email, calendar, contacts, tasks, and OneDrive files via CLI or MCP server, supporting personal and enterprise accounts.

Guides users through Nonviolent Communication to turn emotionally charged thoughts into constructive messages, structuring responses with observation, feeling, need, and request.

Scouts market opportunities, identifies gaps, and surfaces recent demand signals with evidence and next actions via the MSX API.

Coordinates shopping between buyers and merchants: publish products, manage stock, answer questions, compare prices, and create trackable orders with local or registry-backed state.

Publishes long-form articles to hum.pub, handling registration, authentication, SEO metadata, and audience engagement for AI authors.

Builds a personalized fitness system by assessing your body, goals, and constraints, then designs training, nutrition, and recovery plans that adapt to real-life disruptions.

Builds an optimized Ele.me food cart from an authenticated session, comparing real payable totals and respecting budget and preferences before handing off for payment.

Translates raw genomic data into personalized health, longevity, and pharmacogenomic protocols for AI agents.

智能路由中心，自动识别用户意图并路由到相应技能，提供专业化的深度研究体验，支持多Agent并行调研与监控。

Integrates Clock Address Naming with MCP agents to locally stamp, verify, and log tool outputs with hashes and timestamps.

Provides a local-first agent framework for observable execution, composable tool use, and multi-agent orchestration with policy-guarded execution.

Writes and manages professional bios for any platform and purpose, matching your voice and career stage, and keeping them current as your work evolves.

Downloads Douyin videos without watermark, uploads them to Alibaba Cloud OSS, and writes metadata to a Feishu Bitable spreadsheet.

Create, edit, and export videos from the terminal using the Bazaar CLI, with a gated workflow for context, generation, review, and iteration.

Checks on-chain agent trustworthiness, discovers agents for tasks, profiles agents, looks up wallet/identity records, contacts or dispatches agents, and reads or writes hosted metadata via the 8K4 Protocol (ERC-8004).

Extracts YouTube video transcripts and converts them into summaries, chapters, threads, blog posts, and quotes.

Operate Hugging Face Hub repositories, models, datasets, and Spaces via the hf CLI, including downloads, uploads, authentication, and compute jobs.

Manages Apple Reminders from the terminal using remindctl, supporting add, list, complete, and delete operations with due dates and alarms.

Fetches stock quotes, search, history, comparisons, and crypto prices from Yahoo Finance via a read-only Python script.

Query Solana blockchain data with USD pricing — wallet balances, token portfolios with values, transaction details, NFTs, whale detection, and live network stats. Uses Solana RPC + CoinGecko. No API key required.

Search, read, create, update, and delete blocks and documents in a self-hosted SiYuan knowledge base via its kernel API using curl.

Builds production RAG and semantic search systems with Qdrant, covering collection setup, vector indexing, filtered and hybrid search, and integration with LangChain and LlamaIndex.

Configure and troubleshoot Honcho memory for Hermes, including multi-profile peer isolation, observation settings, dialectic reasoning, session summaries, and context budget enforcement.

Store and query embeddings with metadata, vector and full-text search, and filtering. Integrates with LangChain and LlamaIndex for RAG and semantic search applications.

Fetches enrolled courses and assignments from Canvas LMS using API token authentication, with read-only access.

Reads, searches, and sends Telegram messages via a CLI, including inbox summaries, chat history, contacts, and group member lookups.

Downloads videos and music from YouTube, Bilibili, X, and other sites into organized local folders, with optional playlist subdirectories and Telegram delivery.

Debug Python code interactively with pdb and debugpy, covering local breakpoints, post-mortem analysis, and remote attachment to running processes.

Guides LLM post-training with RL using slime, a Megatron+SGLang framework for training GLM, Qwen, DeepSeek, and Llama models with GRPO, async, and multi-turn workflows.

Trains LLMs with SimPO, a reference-free preference optimization method that outperforms DPO, using configurable hyperparameters and workflows for various models and tasks.

Run ML workloads on Modal's serverless GPU cloud: deploy models as auto-scaling APIs, run batch jobs, and schedule tasks with pay-per-second GPU pricing.

Runs the open-source LLaVA vision-language model for image understanding, captioning, visual question answering, and multi-turn image conversations, including setup, inference, and training guidance.

Enables fast similarity search and clustering of dense vectors using FAISS, covering index types, GPU acceleration, and integrations with LangChain and LlamaIndex.

Runs three parallel reviewer agents over recent code changes, aggregates their findings, and applies the fixes worth applying.

Provides detailed guidance on Hermes Kanban worker lifecycle, including workspace handling, handoff patterns, retry diagnostics, and edge cases for dispatched workers.

Create, search, triage, label, and assign GitHub issues using the gh CLI or GitHub REST API.

Guides CLI/API agents through a design process for one-off HTML artifacts like landing pages, prototypes, and decks, emphasizing context gathering and verification.

Fine-tune large language models by training less than 1% of parameters using LoRA, QLoRA, and 25+ adapter methods, enabling efficient adaptation on limited GPU memory.

Delegate coding tasks to xAI's Grok Build CLI, covering headless and interactive modes, PR reviews, and session management.

Enables zero-shot image classification, image-text matching, and cross-modal retrieval using OpenAI's CLIP model, with code for semantic search, content moderation, and vector database integration.

Configure, extend, or contribute to Hermes Agent, an open-source AI agent framework by Nous Research.

Search local knowledge bases, notes, docs, and meeting transcripts with hybrid retrieval combining BM25, vector search, and LLM reranking, all running on-device.

Queries EVM blockchain data across 8 chains with USD pricing, including wallet portfolios, token info, transactions, gas, and contract inspection.

Runs 150+ AI apps in the cloud via the infsh CLI, covering image generation, video creation, search, 3D, and social automation without needing a GPU.

Creates 3Blue1Brown-style animated explainer videos, algorithm visualizations, and math animations using Manim Community Edition.

Sets up GitHub authentication for git and gh CLI, covering HTTPS tokens, SSH keys, and API access.

Delegate coding tasks to the Claude Code CLI via terminal commands, supporting both one-shot and interactive sessions.

Tracks blog and RSS/Atom feed updates using the blogwatcher-cli tool, with feed discovery, HTML scraping fallback, OPML import, and read/unread article management.

Converts video, audio, or images into colored ASCII art videos (MP4/GIF) with generative effects, audio-reactive visuals, and text overlays.