Ai agent builder

Skill aizech/bernhard-zechmann-skills/skills/engineering/ai-agent-builder

Modular AI agent skills used across engineering, writing, and research workflows. Model‑agnostic, composable, and production‑tested. Includes engineering, productivity, personal, and experimental skill packs.

Install

npx -y skills add aizech/bernhard-zechmann-skills --skill ai-agent-builder

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Build an AI agent with tools, memory, and multi-step reasoning. Use when the user wants a chatbot, reasoning agent, or multi-agent system.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

1.8 KB, as published. Nobody here has run it

AI Agent Builder

Design and build an AI agent that uses tools, memory, and reasoning.

Agent types

Reactive: single-turn, no memory. Use for classification or simple Q&A.
Conversational: multi-turn with conversation memory. Use for chatbots.
Tool-using: calls external APIs. Use for actions and lookups.
Reasoning: plans and executes multiple steps. Use for complex tasks.
Multi-agent: specialized agents collaborating. Use for workflows.

Tool design

Clear, action-oriented names.
Zod or Pydantic input schemas with constraints and examples.
Concise descriptions that explain when to use the tool.
Group related tools under a consistent prefix.

Memory patterns

Buffer memory: keep last N messages.
Summary memory: summarize older turns to stay within context.
Vector memory: retrieve relevant facts semantically.
Entity memory: track entities mentioned across the conversation.

Reasoning

ReAct: thought, action, observation, repeat.
Planning: plan first, execute each step, then synthesize.

Always keep the agent's goal visible. Let it ask for missing information before acting.

Output

For a new agent, produce:

Architecture diagram or description.
List of tools with inputs and outputs.
System prompt.
Conversation flow example.
Deployment/integration notes.

Guardrails

Never make up credentials or endpoints.
Escalate to human when confidence is low or stakes are high.
Keep tools read-only until the user explicitly approves writes.