AI Tools & Skills Encyclopedia

schedule — Automated Task Scheduling Skill

A skill that lets AI agents create tasks which can be run on demand or automatically on an interval.

The schedule skill allows an AI agent to move beyond one-off conversations and operate on a clock. With it, an agent can create a task — such as summarising a news feed, checking a database, or emailing a weekly report — and arrange for that task to run at a fixed time, on a recurring interval, or whenever the user triggers it.

How it works

When a user asks the agent to "do X every Monday morning," the schedule skill writes a definition of the task (its prompt, inputs, and cadence) into a scheduler service. The service then wakes the agent on schedule and replays the task, allowing long-running automations without manual intervention.

Typical uses

• Daily digests of email, Slack, or news.
• Weekly financial reports generated from a spreadsheet.
• Hourly monitoring of websites or APIs for changes.
• Reminders tied to natural-language goals, e.g. "every Friday at 4pm, review my to-dos."

setup-cowork — Guided Agent Onboarding Skill

A step-by-step setup experience that helps a new user install a matching plugin, try a skill, and connect external tools.

The setup-cowork skill is a guided onboarding flow. It asks the user about their role and goals, recommends a plugin that matches those needs, walks them through installing it, and then demonstrates a sample skill the user can try immediately. Finally, it prompts the user to connect any external tools (email, calendar, storage, databases) that make the agent more powerful.

Why it matters

AI agents become dramatically more useful once they are connected to a user's own data and workflows. The setup-cowork skill reduces the friction of that first step and prevents users from getting lost in configuration screens.

Steps

• Identify the user's role (developer, analyst, writer, operator…).
• Suggest and install a matching plugin.
• Run a quick "hello world" skill to show value.
• Help connect real tools such as Slack, Drive, or GitHub.

xlsx — Excel Spreadsheet Skill

Comprehensive creation, editing, and analysis of Microsoft Excel (.xlsx) files — including formulas, formatting, charts, and data modelling.

The xlsx skill gives an AI agent full fluency in Microsoft Excel. With it, the agent can create new spreadsheets from scratch, open and edit existing ones, apply formulas, format cells and tables, build charts, and generate full financial models. It is the go-to skill whenever tabular data is involved.

Under the hood

The skill uses specialised Python libraries to read and write the Office Open XML format that underlies .xlsx files. It preserves formatting, supports multi-sheet workbooks, and can round-trip files without corrupting their styles or named ranges.

Example tasks

• Build a 5-year startup financial projection.
• Clean and deduplicate a CSV of 50,000 customer records.
• Create a monthly budget dashboard with charts.
• Add formulas that calculate totals and growth rates automatically.

pdf — PDF Processing Skill

A comprehensive toolkit for extracting text and tables from PDFs, creating new PDFs, merging, splitting, watermarking, and filling forms.

The pdf skill makes an AI agent a complete PDF workshop. It can read text and tables out of existing PDFs (even scanned ones, with OCR), combine multiple documents into one, split a single document into parts, rotate pages, add watermarks, and generate brand-new PDFs from text or HTML source. It can also fill in interactive forms — useful for contracts, applications, and tax documents.

Why PDFs are hard

Unlike HTML or Word documents, PDFs are designed to represent the appearance of a page rather than its semantic content. Extracting structured data (tables, headings, columns) requires specialised parsing. The pdf skill wraps these parsers behind a simple interface so the agent can focus on the user's intent rather than the file format.

Example tasks

• Merge 40 invoices into one PDF for the month.
• Extract a table from an annual report into a spreadsheet.
• Fill in a government form using data from a database.
• Create a polished PDF of a generated report for sharing.

skill-creator — The Meta-Skill

A skill for creating new skills, modifying existing ones, running evaluations, and optimising their descriptions for reliable triggering.

The skill-creator is the "meta-skill": a skill whose only purpose is to produce more skills. Using it, a developer can scaffold a new skill folder, edit its instructions, add supporting scripts, run evaluations to measure how reliably the agent invokes it, and refine its description to improve triggering accuracy.

Why it exists

A skill is only useful if the agent knows when to use it. The skill-creator includes tools for benchmarking — measuring how often an agent correctly selects a skill given different phrasings — and variance analysis to detect flaky descriptions. This turns skill-authoring from guesswork into an engineering practice.

Typical uses

• Create a skill for interacting with a specialised internal API.
• Improve an existing skill's description so it triggers on more paraphrases.
• Run a regression test to prove a skill still works after edits.

pptx — PowerPoint Presentation Skill

Create, read, and edit Microsoft PowerPoint (.pptx) files, including slide decks, pitch decks, templates, layouts, and speaker notes.

The pptx skill gives an AI agent the ability to work natively with PowerPoint files. Whenever a user says "make me a pitch deck" or "turn this report into slides," the agent invokes this skill to create a proper .pptx file — not just a bulleted text outline. It can also read existing decks to summarise their content, edit text, replace images, and merge multiple presentations.

What makes it powerful

Unlike naive approaches that output markdown "slides," the pptx skill produces real .pptx files with layouts, master slides, fonts and themes that look professional when opened in PowerPoint, Keynote, or Google Slides.

Example tasks

• Build a 10-slide investor pitch from a product brief.
• Turn a research paper into a conference talk deck.
• Extract speaker notes from a .pptx and summarise them.

docx — Word Document Skill

Create, read, edit, and format Microsoft Word (.docx) documents with professional touches like tables of contents, headings, page numbers, and letterheads.

The docx skill lets an AI agent behave like a professional document editor. It can generate polished reports, memos, letters and templates; build tables of contents and structured headings; insert images and citations; and perform find-and-replace across long documents. It can even work with tracked changes and comments — the building blocks of collaborative editing.

When to use it

Any task that ends in "…and save it as a Word doc" is a candidate. Because .docx is still the dominant format in business and academia, this skill is one of the most frequently used in the entire catalogue.

Example tasks

• Write a 10-page market research report with TOC and headings.
• Turn meeting notes into a formatted memo with a company letterhead.
• Apply find-and-replace across 30 templates.
• Insert a chart or image into an existing Word document.

cowork-plugin-customizer — Tailoring Plugins to an Organisation

A skill that customises an existing plugin for a specific organisation's tools, connectors and workflows.

Plugins bundle up connectors, skills and commands that turn an AI agent into a specialised worker. The cowork-plugin-customizer skill takes an off-the-shelf plugin and tailors it to an organisation's real-world setup: swapping out generic connectors for the exact SaaS tools the company uses, tweaking skill descriptions to match internal terminology, and adjusting commands to fit local workflows.

Why customise?

A generic plugin may know "project management," but your company might use a specific tracker with custom fields and permissions. Customising the plugin turns a capable agent into a domain expert for your team.

Typical edits

• Replace a generic chat connector with Slack or Teams.
• Rewrite skill descriptions to match internal jargon.
• Add approval steps required by the organisation.

create-cowork-plugin — Building a Plugin from Scratch

A guided, end-to-end skill that helps users design, scaffold, and deliver a brand-new plugin file.

Where the customizer edits existing plugins, the create-cowork-plugin skill helps users invent new ones from nothing. It walks the user through naming the plugin, defining its purpose, choosing which connectors and skills it should include, writing the right triggers and descriptions, and finally packaging everything into a deliverable .plugin file that can be installed elsewhere.

Who is it for?

Anyone who has a repeatable workflow worth bottling up: consultants, internal tool builders, automation engineers, and hobbyists who want to share their best agent recipes with friends or colleagues.

What you walk away with

• A structured plugin folder with skills, commands and configs.
• A packaged .plugin file ready for distribution.
• Clear triggers, so the agent knows when to use it.

Windsurf — The Agent-Native IDE

A sleek, AI-first code editor similar to Cursor, with a clean interface and a powerful agentic workflow.

Windsurf is an AI-first integrated development environment that reimagines what coding feels like when an agent is sitting beside you. Similar in spirit to Cursor, it pairs a clean, distraction-free interface with a deeply integrated agent that can read, navigate, edit, and run your codebase on its own.

Why developers choose it

Windsurf's biggest draw is its agentic flow: instead of asking the AI for a single suggestion at a time, you can hand it multi-step tasks ("add a login page and wire it to our auth service") and watch it plan, edit files, run tests, and iterate. The UI stays out of the way while the agent works.

Best for

• Developers who want an agent that edits across many files at once.
• Teams seeking a minimalist, modern alternative to older IDEs.
• Prototyping and rapid iteration on new features.

Aider — AI Pair-Programming in Your Terminal

A lightweight command-line coding assistant that works directly with Git.

Aider is an open-source AI pair programmer that runs entirely in your terminal and integrates tightly with Git. Instead of wrapping an editor around an LLM, Aider treats your repository as the source of truth: every change the AI makes becomes a real commit, fully visible and revertible with standard Git tools.

Why it stands out

Because Aider is lightweight and Git-native, it fits naturally into any existing workflow. It doesn't ask you to switch IDEs, adopt a new UI, or trust an opaque change-tracking system — the history is just your commit log. This makes it especially popular with experienced developers who want AI help without giving up control.

Best for

• Terminal-first developers and open-source contributors.
• Anyone who wants every AI edit captured as a proper Git commit.
• Working on large legacy codebases where transparency matters.

Cline — Autonomous AI Agents Inside VS Code

An extension that runs full AI agents directly inside Visual Studio Code.

Cline brings autonomous AI agents directly into the editor millions of developers already use: Visual Studio Code. Installed as an extension, it gives VS Code an agent that can read your project, make changes across files, execute terminal commands, and iterate toward a goal — all without leaving your familiar editor.

Why it's interesting

Rather than forcing developers to adopt a new IDE, Cline meets them where they already are. This lowers the friction of trying agentic coding and lets users keep all their existing themes, keybindings, and extensions.

Best for

• Developers committed to VS Code who want agentic power-ups.
• Teams that need a free, open ecosystem for AI-assisted coding.
• Automation tasks that span editing files and running shell commands.

Smart Connections — Semantic Linking for Your Notes

An Obsidian plugin that uses embeddings to surface related notes and enable natural-language chat with your vault.

Smart Connections is an Obsidian community plugin that indexes every note in your vault using vector embeddings, then uses those embeddings to surface related notes as you write. You can also chat with your whole vault in natural language, asking questions that pull from dozens of notes at once.

Why it matters

Traditional note apps rely on manual links and folders. Smart Connections adds a second, automatic layer of structure based on meaning rather than keywords — turning a pile of notes into a living second brain that can answer questions.

Text Generator — LLM Writing Inside Obsidian

A powerful plugin that lets Obsidian users call any LLM to draft, continue, summarise, or transform text.

The Text Generator plugin brings powerful LLM writing capabilities to Obsidian. Users can highlight text and ask the AI to continue it, rewrite it, summarise it, or generate brand-new content using custom prompt templates stored as notes in the vault. It supports many LLM providers, making it model-agnostic.

Why it stands out

Its "templates as notes" design is uniquely Obsidian: your prompts live alongside your content and can be version-controlled and shared like any other note. This turns prompt-engineering into a first-class part of your knowledge base.

Copilot (for Obsidian) — An AI Chat Sidekick in Your Vault

A community plugin that adds a ChatGPT-style chat pane to Obsidian, scoped to the notes you care about.

The Copilot plugin for Obsidian adds a ChatGPT-like chat pane to the side of your editor. You can ask it questions about the current note, the entire vault, or unrelated topics. It supports multiple models and lets you save useful chats as new notes in the vault.

Why it matters

It is the simplest way to add a conversational layer on top of an existing Obsidian vault. Together with Smart Connections and Text Generator, it forms the "AI triad" many Obsidian power users rely on.

Notion AI — Docs, Databases and AI in One Workspace

Notion's built-in AI assistant combines documents, databases and generative writing into a single workflow-friendly platform.

Notion AI is the AI layer built directly into the popular Notion workspace. Because Notion already blends documents, wikis, and databases, its AI can draft documents, summarise meeting notes, auto-fill database fields, translate content, and extract action items — all within the same place your team already works.

Why it matters

Notion AI is especially good for workflows: use a database to track projects, then let the AI summarise progress, generate status reports, or draft emails based on rows in that database. It turns Notion from a filing cabinet into an active teammate.

Mem.ai — Notes That Organise Themselves

A self-organising note-taking app that uses AI to connect and surface your thoughts — like a brain.

Mem.ai is a note-taking app built around a simple idea: you shouldn't have to organise your notes — the AI should. Just write, and Mem uses AI to tag, link, cluster and surface relevant past notes without the need for folders or manual filing.

Why it stands out

Mem is often described as "working like a brain." Instead of forcing rigid structure, it lets ideas flow in, then quietly builds connections behind the scenes and resurfaces them when they become relevant again. It's ideal for people who collect thoughts faster than they can file them.

Reflect Notes — A Daily Thinking System with GPT Built In

A beautifully designed note-taking app centred around daily notes, backlinks, and a native GPT assistant.

Reflect Notes is a minimalist note app built around the idea of daily notes — a fresh page for every day, with powerful backlinks that connect your thoughts over time. Its built-in GPT assistant can summarise, draft, brainstorm and query your notes on demand.

Why it stands out

Reflect pairs the beauty of apps like Bear with the graph-thinking of Roam Research and the power of built-in GPT. It's designed to be the single place where your daily thinking lives — with AI woven in as a natural collaborator.

Logseq — The Local-First Knowledge Graph

An open-source outliner and knowledge graph that stores your data locally and offers a powerful graph view.

Logseq is an open-source, local-first outliner and knowledge graph. All of your notes are stored as plain markdown files on your own machine, meaning you fully own your data and can edit it with any text editor. Its standout feature is a powerful, interactive graph view that visualises how your notes connect.

Why it stands out

For users who don't want their knowledge trapped in a cloud silo, Logseq offers the best of both worlds: the fluid bi-directional linking of modern PKM apps, combined with the permanence and portability of plain files.

n8n — Visual Workflow Automation for APIs and AI

An open-source, node-based automation platform that lets you connect APIs, databases and AI models into powerful workflows.

n8n is a visual, node-based workflow automation tool that lets you connect hundreds of services — APIs, databases, SaaS tools, and AI models — without writing much code. It's often described as a more flexible, self-hostable alternative to Zapier or Make.

Why it matters for AI

With the recent explosion of LLM nodes, n8n has become one of the easiest ways to build real AI pipelines: pull data from an API, run it through GPT or Claude, transform the output, and push it to Slack, a database, or a file. All visually, and all self-hostable.

OpenRouter — One API for Every AI Model

A unified gateway that gives developers access to dozens of LLMs through a single API and routes tasks dynamically.

OpenRouter provides a single API endpoint that can call any of dozens of LLMs from different providers — OpenAI, Anthropic, Google, Meta, Mistral, and many more. Developers can switch models with one line of code, compare pricing and performance, and even route different tasks to different models dynamically.

Why it matters

In a world with hundreds of LLMs, locking yourself into one provider is risky. OpenRouter gives teams the flexibility to choose the best model for each job — a fast model for classification, a deep-thinking model for reasoning — all without rewriting their code.

LangChain — The Framework for Custom AI Agents

A widely used open-source framework for building LLM-powered applications and custom agents.

LangChain is one of the most popular open-source frameworks for building applications powered by large language models. It provides reusable building blocks for common AI patterns — prompt templates, memory, tool use, retrieval-augmented generation (RAG), and multi-step agents — so developers don't have to reinvent the wheel.

Why it matters

If you want to build a custom AI agent that combines an LLM with databases, APIs, search tools and memory, LangChain gives you a standard vocabulary and set of components. It is one of the foundations of the modern agentic AI ecosystem.

AutoGen — Multiple AI Agents Working Together

A Microsoft-backed open-source framework for creating conversations between multiple LLM agents that cooperate to solve tasks.

AutoGen is an open-source framework from Microsoft Research that treats AI applications as conversations between multiple specialised agents. Rather than asking one giant LLM to do everything, you define several agents — a planner, a coder, a critic, a user proxy — and let them talk to each other until they solve the task together.

Why it matters

Multi-agent systems often outperform single agents on complex problems because each agent can focus on a single role. AutoGen makes this pattern practical, with built-in support for code execution, tool use, and human-in-the-loop oversight.

OpenClaw — Full System Automation That Executes Real Tasks

An agentic automation platform capable of controlling a full computer environment to execute real-world tasks end-to-end.

OpenClaw is an agentic automation platform designed for end-to-end execution of real-world tasks. Instead of generating a plan for a human to follow, OpenClaw actually performs the plan itself — controlling files, applications, web browsers, and even system-level operations when needed.

Why it matters

Most AI tools stop at suggesting what to do. Systems like OpenClaw close the loop by taking action, turning conversational AI into genuine digital labour. This raises the bar for what an AI assistant can deliver: not advice, but completed work.

Common uses

• Running repetitive multi-app workflows on your computer.
• Automating back-office operations with no custom scripts.
• Executing a task described in plain English from start to finish.

ElevenLabs — Realistic AI Voice and Audio

A leading AI audio platform known for hyper-realistic text-to-speech, voice cloning, and multilingual voice generation.

ElevenLabs is widely regarded as one of the best AI voice platforms in the world. Its text-to-speech engine produces voices that are almost indistinguishable from human recordings, complete with natural breaths, pauses, and emotional inflection. It also offers voice cloning, which can recreate a speaker's voice from a short sample, and multilingual output that can speak over thirty languages while preserving the speaker's identity.

Why it matters

Realistic voice has unlocked a new generation of products: AI narrators for audiobooks, voice-overs for videos, accessibility tools for the visually impaired, and character voices for games and interactive media. ElevenLabs is the default choice for builders who need production-quality audio without a studio.

Best for

• Podcasters, YouTubers, and video producers generating narration.
• Indie game developers giving characters real voices.
• Accessibility products that read text aloud in a natural voice.
• Localising content into new languages using the same speaker.

Suno — AI Music Generation from a Prompt

A generative music platform that turns a text description into full songs complete with instruments, vocals and lyrics.

Suno is a generative music platform that can produce fully-arranged songs — complete with melody, instrumentation, vocals, and lyrics — from a simple text description. Users can ask for "an upbeat indie pop song about summer road trips" and get back a complete track in under a minute. Advanced controls let creators tweak style, tempo, instrumentation and lyrics.

Why it matters

Music has traditionally been one of the hardest media to create without training or equipment. Suno collapses that barrier, letting anyone prototype a song, create background music for a project, or iterate on melodies without ever touching an instrument.

Best for

• Generating original background music for videos and games.
• Songwriters prototyping melodies and arrangements.
• Social media creators producing unique, royalty-free tracks.

Runway — AI Video Generation and Editing

A creative suite that lets users generate, edit, and transform video using a growing collection of AI models.

Runway is the leading creative platform for AI video. Its Gen series models can generate new video clips from a text prompt, an image, or an existing clip, and the platform also includes a full suite of AI-powered video editing tools: rotoscoping, background removal, inpainting, motion tracking, and more.

Why it matters

Runway turns what used to be hours of painstaking post-production into a few clicks. Independent filmmakers and small studios can now produce visual effects shots, stylised sequences and concept videos that previously required entire VFX teams.

Best for

• Generating short cinematic clips from text or images.
• Removing backgrounds and objects from footage automatically.
• Prototyping ads, music videos and concept films quickly.

Midjourney — The Iconic AI Image Generator

One of the most recognisable AI image generation tools, known for its stunning artistic style and active creative community.

Midjourney is one of the most famous AI image generation tools, widely praised for the beauty and artistry of its output. Users describe an image in text — a "cyberpunk cityscape at dusk" or "a still life in the style of Vermeer" — and Midjourney returns four variations, which can then be refined, upscaled, or remixed.

Why it stands out

Where many image generators aim for photorealism, Midjourney leans into a distinctive cinematic and painterly aesthetic that has made it a favourite of concept artists, brand designers and marketers. Its tightly knit community shares prompts and techniques, making it easy to learn from others.

Best for

• Concept art, moodboards and storyboards.
• Brand and marketing imagery with a distinctive style.
• Rapid visual ideation for product and game design.

Perplexity AI — The Answer Engine

A conversational AI search engine that answers questions with citations to real sources, designed to replace Google for research.

Perplexity AI calls itself an "answer engine" rather than a search engine. Instead of returning a list of links, it reads across the web in real time and returns a direct, conversational answer — always backed by numbered citations to the sources it used. Users can click through to verify, ask follow-up questions, or drill deeper into a topic.

Why it matters

For research-heavy work, Perplexity dramatically cuts the time between question and insight. Instead of opening a dozen tabs and reading each one, you get a synthesised answer with sources, and you can verify the parts that matter. Its Pro and Research modes can tackle deeper, multi-step queries with longer reasoning.

Best for

• Academic and professional research with source-level citations.
• Fast fact-finding when accuracy matters.
• Comparing multiple perspectives on a current topic.

You.com — Customisable AI Search and Assistants

An AI-first search engine that combines chat, traditional search results, and customisable AI assistants in one place.

You.com is an AI-first search platform that blends a traditional search engine with conversational AI. Users can choose from several models, run deep research tasks, get cited answers, and even access specialised "apps" for things like coding, writing or academic research. It is designed to be configurable: each user can customise what kinds of results they see and which AI models handle their queries.

Why it matters

You.com positions itself as a privacy-friendlier, customisable alternative to Google and Bing, with much deeper AI integration. Its ability to switch between fast answers and longer research modes makes it a flexible daily driver for knowledge workers.

Best for

• Users who want both conventional search and AI chat in one tab.
• Running side-by-side comparisons across multiple LLMs.
• Research workflows that need both quick facts and deep dives.

NotebookLM — Your Personal Research Notebook by Google

An AI-powered research tool from Google that grounds its answers in the documents you provide, then helps you summarise, explore and even create audio overviews.

NotebookLM is Google's AI-powered research notebook. Unlike a general chatbot, NotebookLM is grounded in the sources you upload — PDFs, Google Docs, articles, or slides. Its answers, summaries and outlines all cite the exact passages from your own documents, making it ideal for academic and professional research where trust and traceability matter.

Standout features

Beyond question answering, NotebookLM can generate study guides, timelines, FAQ lists, and even full "audio overviews" — conversational podcast-style summaries of your sources — making it easy to absorb complex material on the go.

Best for

• Students synthesising textbooks, papers and lecture notes.
• Analysts producing briefings grounded in trusted sources.
• Anyone who wants an AI assistant that cannot hallucinate beyond the documents provided.