- Home
- Skills
- Prompt & Skill Engineering
- Agent Engineering Kit
Works with the AI Agents you already use
Agent Engineering Kit
Ask for a focused result: - A feature design with clear responsibilities, contracts and acceptance criteria.
Free
Agent Engineering Kit
Example session with this skill installed
Use Agent Engineering Kit to recommend profiles for a browser-only React + TypeScript app. We are not using Next.js, a Node.js backend or a database. We already have AGENTS.md and project checks. Prepare a recommendation only: do not edit files, install anything or run commands.
- Read your context and instructions
- Compiled the agent engineering kit
Recommendation
Select the react and typescript profiles. Their dependencies make the shared JavaScript and engineering guidance available. Keep the existing AGENTS.md and required project checks until a concrete adoption proposal has been reviewed.
This is a recommendation based on the stack stated in your request. No repository inspection, commands, installation or file changes were performed for this example request.
| Guidance | Why it applies |
|---|---|
| React | Component and Hook boundaries, state ownership, effects, forms and accessible interaction |
| TypeScript | Explicit typed contracts, validation boundaries and the project's strict analyzer configuration |
| JavaScript | Shared language behavior, async effects and module structure |
| Shared AEK rules | Responsibility placement, task scope, proportionate checks and accurate completion evidence |
Keep selection scoped
Do not select Next.js, Node.js server, database, ORM or container guidance merely because this is a React application. A local Node-based build tool would be a separate tooling concern, not evidence of a Node.js backend. Profile dependencies keep reference files available; they do not require reading every workflow or setting up delegation.
Information needed for an adoption proposal
Provide the target project path and chosen coding client, plus access to the current AGENTS.md, package manifests and check configuration. Confirm the supported React and TypeScript versions from that project. Preserve its existing mandatory checks and any accepted local contracts.
The next deliverable would identify the selected files, proposed instruction changes, conflicts and verification commands. It would remain a preview until application is authorized. No version upgrade or model change is implied.
Connects securely to your tools. The creator never sees your data.
About this skill
Engineering guidance for the project you actually have
Agent Engineering Kit (AEK), by Vitalii Khomiakov, helps an AI coding agent apply consistent architecture, implementation and review practices to your stack. Ask for a focused result:
- A feature design with clear responsibilities, contracts and acceptance criteria.
- A code review with prioritized findings, concrete failure scenarios and evidence.
- Guidance during authorized implementation, using the relevant technology rules.
- A recommendation of project rules, followed by a reviewable adoption proposal when you want persistent integration.
Provide the relevant code or project context and your intended outcome. The agent selects the applicable references and explains missing context and checks it has not run. Editing and execution depend on the tools and permissions you provide.
Try a review without commands or edits
Provide your code with this request:
Use Agent Engineering Kit to review the code I provide. Identify concrete defects and architecture risks, prioritize the findings, and explain the evidence for each. Report missing context and checks you did not run. Do not edit files or run commands.
Expect findings tied to specific code and failure scenarios, with severity and coverage limits. If no code is supplied, the agent must request it. This is a usage example, not a claim that a live review test has been executed.
Two ways to use AEK
Use through Agensi: connect your agent to Agensi and ask it to use Agent Engineering Kit on a specific task. This edition supplies a dated reference snapshot loaded on demand. It does not install project instructions.
Integrate into a project: obtain the source repository separately and ask your agent to read its SETUP.md. The wizard inspects or asks about your stack, proposes applicable instructions and follows MIGRATION.md for authorized application and updates.
Coverage
The 2026-11-02 catalog has 44 profiles, including shared workflows and compatibility entries. Coverage includes Python, JavaScript/TypeScript, PHP, Go, PostgreSQL, Docker, RAG and LLM applications, Unreal Engine and Godot. Framework guidance includes React, Next.js, Angular, FastAPI, Symfony, Gin and others listed in the source catalog. Select only what applies to your project.
Requirements and limits
Use an agent connected to Agensi that can read text references, plus the relevant project context or supplied code. Reviewing text supplied in the conversation needs no local file access, terminal, Python, Git, Node.js or additional MCP server. Repository inspection needs access to relevant files; authorized edits, tests, external research and specialized integrations need the tools for that task. Available technology guidance does not make its tools universal prerequisites.
AEK targets Codex, Claude Code and Cursor workflows. Other MCP-capable agents may read the references; native instruction loading and configuration need separate verification. Superpowers is a separate plugin and is not included. Examples retain their version conditions. AEK does not upgrade dependencies automatically or guarantee agent behavior.
Hosted reference snapshot: 2026-11-02. Check the source repository for current contents; an existing project adoption is updated separately.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.
The hosted edition is a dated reference snapshot. Review coverage depends on supplied context; missing context and unrun checks must be reported. Local implementation, verification and persistent adoption require their own tools and authorization. Examples retain version conditions. Native client activation and model behavior need separate verification. Superpowers is not bundled. Source changes do not automatically update this hosted edition.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 14 days ago
- Free with an Agensi account
- Accounts
Python Httpx
- Software
Python
Node.jsEditor MCPTooling MCPUnreal Engine MCPUnreal MCP
Git
Permissions
Allowed Hosts