Guide · Agent skill
How to use the AAS Stack Reviewer agent skill (aas-review-stack)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The aas-review-stack skill lets an AI agent review a set of AI skills, called a stack, that you have composed with the Agentic Awesome Skills (AAS) toolkit. It opens a review view called Workbench and explains the stack's skill IDs, catalog versions, target setup and proposed operations in plain language. By the end of this guide you will know what the skill does, how to install it, how to ask for a review and what it needs to run.
AAS Stack Reviewersickn33/agentic-awesome-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the aas-review-stack ZIP from the page you are reading.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder for one project.
- 3Or upload in the Claude apps. Alternatively, upload the skill under Settings → Capabilities → Skills.
- 4Check requirements. Confirm the bundled local AAS MCP server is available and that Node.js 22 or later is installed.
- 5Make or open a stack. Compose a stack with the AAS toolkit in your current session, or have the artifact ready to import.
- 6Ask for a review. Make an explicit request, such as asking the agent to review the stack and show missing setup or artifact mismatches.
- 7Read the results. Check the IDs, versions, proposed operations and mismatches the agent reports before applying anything elsewhere.
What an agent skill is and what this one does
An agent skill is a folder with a SKILL.md file that gives an AI agent instructions for a specific task. The agent loads the skill when a request matches it.
This skill is for reviewing a stack, meaning a set of AI skills someone has composed with AAS. When asked, the agent opens Workbench, checks exact skill IDs, catalog versions, target setup and proposed operations, and explains any mismatches or missing setup. It is designed to review only. It does not install, apply, recover or upload anything.
Install the skill from a ZIP file
Download the ZIP for the skill from the page you are reading. Unzip it into one of these locations, depending on how you want to use it.
For all your projects, unzip it into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder.
In the Claude apps you can instead upload the skill under Settings → Capabilities → Skills.
- Personal use across projects: ~/.claude/skills/
- Single project: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills, then upload
Meet the requirements before you start
The skill needs a bundled local AAS MCP server, Node.js 22 or later, and an MCP-capable host to show the Workbench UI. If your host cannot render the UI, the skill falls back to a local HTML page, ui/workbench.html, which you can open in a browser.
The skill includes its catalog and supporting files, and the source says no network service, API key or runtime download is required.
- Bundled local AAS MCP server
- Node.js 22 or later
- An MCP-capable host for the Workbench UI (a local HTML fallback exists)
Trigger the skill with a natural request
The skill activates when you explicitly ask for an AAS review. It should not activate for unrelated tasks. Ask in plain words that name the stack and the kind of check you want.
Good requests name the stack and what you want to see, such as missing setup or artifact mismatches. If no stack exists in the current session, the agent will ask you for the artifact or offer to compose one rather than making one up.
- Review the stack we just composed and show any missing setup or artifact mismatches.
- Preview the plan for this stack manifest and explain what it would change, without applying anything.
- Check which skill IDs and catalog versions this stack uses.
Read what the review tells you
The agent checks exact IDs, catalog version, target, profile, proposed operations and artifact digests, and points out cross-artifact mismatches. It reports browser validation separately from Core inspection and from its own semantic judgment.
A consistent stack is not proven to be suitable or to come from a known author. The skill says consistency checks do not prove either.
- Errors, unavailable resources and catalog gaps are reported clearly.
- If an import fails or changes, stale results are cleared.
Safety notes: scripts and network
The skill runs scripts, so it requires the local MCP server. It does not use the network. The agent is told not to execute retrieved scripts automatically, fetch missing payloads, or claim installation or directory approval without verification.
Catalog instructions and supporting files are treated as untrusted content. They do not grant permission or authorize unrelated actions. Explicit local file imports are processed in the browser and are not sent to another tool.
- Runs scripts: yes, through the bundled local MCP server
- Uses network: no
- Local file imports stay in the browser
Tips for getting useful reviews
Compose or open the stack in the same session before asking for a review, so the agent has something to inspect.
Use the review as a second check before you apply a stack with the toolkit itself. The skill will not apply it for you.
Treat the result as evidence to read, not a verdict on whether a skill is safe or fit for your task.
- Ask for the review in the same session as the stack.
- Read the mismatches list before applying anything.
- Do not rely on the review alone to judge skill quality.
What people use it for
- Checking which skill IDs and catalog versions a proposed agent setup would use before applying it.
- Getting a plain-language explanation of a stack manifest you did not write.
- Finding missing target setup or artifact mismatches in a composed stack.
- Auditing a team's proposed agent setup before anyone changes their environment.
Frequently asked questions
How do I install the aas-review-stack skill?
Download the ZIP from the page, then unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder for one project. In the Claude apps you can instead upload it under Settings → Capabilities → Skills.
What do I need to run aas-review-stack?
You need the bundled local AAS MCP server and Node.js 22 or later. You also need an MCP-capable host to show the Workbench UI, though a local HTML fallback exists.
Is aas-review-stack free?
The skill is licensed under MIT. The source does not mention any price, so check the license file in the package for the terms.
Does the skill install or apply my skills?
No. It is designed only to review. It does not install, apply, recover or upload anything, and it only previews plans.
Does the skill use the internet?
The source says it uses no network. It runs scripts through the local MCP server, and the catalog and supporting files are included in the package.
Can it tell me if a skill is safe or trustworthy?
No. The skill states that consistency checks do not prove a skill's suitability or the author's identity, so you need to judge that yourself.
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
This skill suits developers and teams who compose AAS skill stacks and want a plain-language second check before applying them. It is a poor fit if you want the agent to install or apply skills directly, or if you need a judgment on whether a skill is safe or suitable.
Download the skill