Aman FAQs

Question 1

What is Aman and its primary function?

Accepted Answer

Aman is an MCP (Model Context Protocol) server designed to provide AI agents with a standardized, programmatic interface to access and manage their core operational elements: identity, tools, workflows, guardrails, and evaluation within the Aman ecosystem.

Question 2

Is Aman compatible with popular AI agent platforms or LLMs?

Accepted Answer

Yes, Aman speaks the standard Model Context Protocol over stdio, making it compatible with any MCP client. Specific setup instructions are provided for popular AI agent platforms like Claude, Cursor, and Windsurf, enabling easy integration.

Question 3

What specific aspects of an AI agent can be managed through Aman?

Accepted Answer

Aman offers control over: `acore` (identity and personality), `akit` (tool registry and search), `aflow` (reusable workflows), `arules` (guardrails and safety rules), and `aeval` (evaluation metrics and session logging).

Question 4

What is the broader Aman ecosystem?

Accepted Answer

The Aman ecosystem is a comprehensive suite of layered components (acore, amem, akit, aflow, arules, aeval) that collectively define and manage an AI's identity, memory, capabilities, and behavior. Aman-mcp acts as the central bridge, unifying programmatic access to all these layers.

Question 5

How does Aman enable AI agents to manage their functionality?

Accepted Answer

Aman exposes various functions as MCP tools across five layers (`acore`, `akit`, `aflow`, `arules`, `aeval`). This allows AI agents to programmatically read and write their identity and personality, discover and utilize tools, manage reusable workflows, enforce safety guardrails, and log/retrieve performance metrics.

Aman

Aman

主要功能

使用案例

主要功能

使用案例