How to use the AgentFolio Agent Discovery skill

What is an agent skill and what does this one do?

An agent skill is a folder containing a SKILL.md file with instructions that an AI agent reads when a task matches the skill's description. This skill is named agentfolio and is described as a skill for discovering and researching autonomous AI agents, tools and ecosystems using the AgentFolio directory.

Once installed, the agent knows how to act as a guide for finding existing agents and frameworks. It suggests search terms, lists what to note about each candidate, and helps turn findings into comparison tables or build-versus-buy notes.

• Suggests search terms based on the problem you want to solve
• Lists what to capture for each candidate: core promise, inputs and outputs, autonomy model and deployment model
• Helps build comparison tables and build-versus-buy notes
• Does not pull live data automatically: it relies on the agent or you visiting the website

How do you install the agentfolio skill?

Download the ZIP of the skill from this directory page and unzip it. You can install it in one of two places, depending on whether you want it available everywhere or only in one project.

For a personal install across projects, place the unzipped skill folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.

In the Claude apps you can instead upload the skill under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/
• Project install: .claude/skills/ inside your project
• Claude apps: Settings → Capabilities → Skills, then upload

How do you trigger the skill with a natural request?

You do not need a special command. Ask in plain language about discovering, comparing or researching AI agents, and the agent can apply the skill when the request matches its description. Naming AgentFolio in your request makes it more likely to be used.

Good requests state the problem, the category and the output you want, for example a list of options with specific attributes to compare.

• Use AgentFolio to find 3 autonomous AI agents focused on code review, and summarize the core value prop and supported languages.
• Scan AgentFolio for customer support triage agents and list notable UX patterns.
• Before we build our own research assistant, map existing research agents and highlight gaps we could fill.

How do you run a landscape scan step by step?

Start by opening https://agentfolio.io in your browser and optionally filtering by category such as Dev Tools, Ops, Marketing or Productivity. Then search by the problem you want to solve, such as customer support agents, autonomous coding agents or research and analysis agents.

For each candidate, capture four things: the core promise, the input and output shape (APIs, UI, data sources), the autonomy model (one-shot, multi-step, tool-using or human-in-the-loop) and the deployment model (SaaS, self-hosted, browser, IDE and so on). Then decide whether to integrate an existing agent or build your own.

• Define the problem in one sentence
• Search the directory using intent-based keywords
• Record the four attributes for each relevant entry
• Mark anything the listing does not state as missing

How do you check the results before relying on them?

Treat directory entries as leads, not proof. The skill itself warns that entries can be stale or promotional, so record when you accessed each page and separate vendor claims from independent evidence.

Verify current pricing, licensing and advertised benchmarks against each vendor's primary documentation before you recommend adoption. The skill is not a substitute for testing or expert review.

• Check pricing and licence terms on the vendor's own documentation
• Treat benchmarks as vendor claims unless independently verified
• Run a proof-of-concept before committing to a vendor

What are the safety notes and requirements?

The skill needs web browsing access to agentfolio.io, because it directs you or the agent to visit the site. It runs no scripts and contains no code or data feed. It does use the network, since browsing is the way it gathers information.

The skill is intended for scoping and research. Use it only when your task matches its scope, and ask for clarification when inputs, permissions or success criteria are missing.

• Requires web browsing access to agentfolio.io
• Runs scripts: no
• Uses network: yes (through browsing)
• Not suited to verifying vendor claims, pricing or benchmarks, or to tasks needing guaranteed-accurate live data