Guide · Agent skill
How to use the Aident Loadout agent skill (aident-loadout)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The aident-loadout skill is a set of instructions that teaches an AI agent how to use Aident Loadout, a hosted service that connects your everyday apps once so the agent can search for, check and run actions in them. This guide explains what the skill changes, how to install it and connect the Aident server, how to ask for things in plain language, and what to watch for. By the end you will be able to set up the skill, run a multi-app task, and understand why the agent stops for your approval before anything changes.
Aident Loadoutsickn33/agentic-awesome-skills · MIT Download the skill Quick start
- 1Create an Aident account. Sign up at https://aident.ai. The tools will not work without an account.
- 2Install the skill. Download the ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings, Capabilities, Skills in the Claude apps.
- 3Add the aident MCP server. In Claude Code, run claude mcp add --transport http aident https://loadout.aident.ai/mcp. Other MCP clients can add the same URL as a Streamable HTTP server.
- 4Sign in with OAuth. The first time you use a service, sign in through the browser. Never paste API keys or passwords into chat.
- 5Make a natural request. Describe the app and the outcome you want, for example a Linear summary posted to Slack.
- 6Approve the preview. Read the target, effect and inputs the agent shows, then reply yes only if they are correct.
What the aident-loadout skill makes the agent do
An agent skill is a folder with a SKILL.md file that tells an AI agent how to handle a certain kind of task. This one covers Aident Loadout. The user connects apps such as Gmail, Slack, GitHub, Linear, Notion or Google Sheets through OAuth in the browser, and API keys are kept in Aident Vault. Any MCP client can then search those apps' actions by intent, read the live schema, check cost and run the action.
The skill sets the order of operations. The agent looks up the action, reads its current schema, checks that the account is connected, checks cost where needed, and states the target and effect of any change before running it. Every run is recorded in Audit on the Aident side. The tools themselves come from the hosted aident MCP server, not from the skill file.
Install the aident-loadout skill
Download the ZIP from the aident-loadout page on this directory, then choose one of two routes. For a personal setup, unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. In the Claude apps, you can instead upload the ZIP under Settings, then Capabilities, then Skills.
The skill package contains one file, SKILL.md. The skill is only instructions, so it does not run any scripts by itself. It does use the network, because the tools it points to are hosted.
- Personal: unzip into ~/.claude/skills/
- Single project: unzip into .claude/skills/ inside the project
- Claude apps: Settings, then Capabilities, then Skills, then upload the ZIP
Connect the Aident MCP server
The skill needs the hosted aident MCP server. You also need an Aident account, created at https://aident.ai. If the aident tools are missing in your session, add the server. In Claude Code, run the command below. Then sign in with OAuth in the browser the first time you use it.
For other MCP clients, add a Streamable HTTP server at the same address and sign in with OAuth. Aident's documentation is at https://docs.aident.ai/loadout/overview. The skill tells the agent never to ask you for credentials, API keys, MFA codes or other secrets in chat.
claude mcp add --transport http aident https://loadout.aident.ai/mcpTrigger the skill with natural requests
You do not need a special command. Ask for something that involves a service you have connected through Aident Loadout, and the agent should pick up the skill. Name the app and the outcome. Requests about which Aident action can do a task, what it will cost, or whether an account is connected also fit.
The skill is not meant for questions the conversation can already answer, or for services you have not connected. If you want a reusable multi-step procedure rather than one action, the skill can also search Aident's Skills library for domain guidance.
- Find the open Linear issues assigned to me and post a summary in #standup on Slack
- How much would it cost to scrape these 20 pages?
- Is my GitHub account connected to Aident?
Review the approval step before anything changes
Before any action that changes external data, the agent must state the target, the effect and the key inputs, then wait for your acknowledgement. Sending a message, publishing, deleting or disconnecting all fall under this rule. If the agent shows you a Slack channel and message, read it carefully. Only say yes when the content is what you want.
If a service is not connected, the agent should give you the connection link and stop. If an action costs credits, it should report the cost and ask before running. The agent is also told not to claim success until the result confirms it.
- Check the channel, recipient and message text in the preview
- Confirm any cost the agent reports before it runs
- Expect the agent to stop if a connection is missing
Safety notes: scripts, network and data
The skill itself runs no scripts and has no install step beyond copying the files. It does use the network, and the actions it runs reach real accounts in the services you connected. The hosted Aident service is closed source, while the skill file is MIT-licensed.
Use the skill for work you would be comfortable doing yourself with approval. Avoid sensitive or irreversible tasks, even with an approval step. Keep provider access within your own authorized accounts.
- No scripts run from the skill package
- Network use is required because tools are hosted at loadout.aident.ai
- Built-in and metered services draw from your Aident balance
- Action schemas change, so the agent reads them fresh each time
Tips for getting good results
Name the service and the end result in your first message. This helps the agent search for the right action. Ask for a cost check before large batch jobs. Keep the connection status in mind, since a missing connection will stop a run early.
If an action fails or the agent cannot find a tool, confirm that the aident server is added and that you signed in with OAuth for that service.
- Be specific about the app, the object and the outcome
- Ask for a preflight cost on anything metered
- Re-run the connection check if a service was recently disconnected
What people use it for
- Read open issues in Linear and post a summary to a Slack channel after you approve the message
- Check what a web scraping job would cost across a list of pages before running it
- Confirm whether a service such as GitHub or Notion is connected to your Aident account
- Find the Aident action that matches a task, such as listing Google Sheets rows
- Run a multi-step research or outreach procedure from Aident's Skills library
Frequently asked questions
Is the aident-loadout skill free?
The skill file is MIT-licensed. The hosted Aident Loadout service is closed source, and built-in and metered services draw from your Aident balance, so some actions have a cost. The skill's preflight step reports the cost before a metered action runs.
What do I need before I can use aident-loadout?
You need an Aident account at https://aident.ai, the hosted aident MCP server at https://loadout.aident.ai/mcp added to an MCP client such as Claude Code, and OAuth sign-in in the browser for each service you connect. An Aident balance is needed for metered or built-in services.
Where do I install the aident-loadout skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps you can upload the ZIP under Settings, then Capabilities, then Skills.
Does the skill run scripts on my computer?
No. The skill package contains only SKILL.md and runs no scripts. It does use the network, since the tools it relies on are hosted by Aident.
Why does the agent keep asking for approval?
The skill requires the agent to state the target, effect and inputs of any action that changes external data, and to wait for your acknowledgement. This is a hard gate, so the agent will not send, publish, delete or disconnect without it.
How is this different from connecting an app directly to my agent?
Aident Loadout connects many services once and lets any MCP client search and run their actions, with cost checks, approval gates and an audit log. The skill gives the agent the procedure for using it. A single direct integration may be simpler if you only need one app and no audit trail.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use aident-loadout if you already have an Aident account and want your agent to read or update the apps you connected, with a cost check and your approval before any change, plus an audit trail. Pick something else if you do not want to create an Aident account, if the task is sensitive or irreversible, or if a normal conversation can answer your question without connecting any services.
Download the skill