skills.rest
michaelalber
michaelalberCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-spec-writer

Generates AI agent blueprints with specified workflows and constraints.

2|Updated Jan 31, 2026
Search Tags:#prompt engineering#specification#ai development#agent design#prd#guardrails
One-click install
npx skills add https://github.com/michaelalber/ai-toolkit --skill agent-spec-writer-michaelalber
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-spec-writer
Source: https://github.com/michaelalber/ai-toolkit/tree/main/skills/agent-spec-writer
Command: npx skills add https://github.com/michaelalber/ai-toolkit --skill agent-spec-writer-michaelalber

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill guides users through the process of designing robust and reliable specifications for AI agents, ensuring clarity, verifiability, and safety.

Core Features & Use Cases

  • Structured Spec Design: Follows a phased approach (Orient, Vision, Structure, Guardrails, Validate, Generate) to build comprehensive agent specifications.
  • Boundary Definition: Enforces the creation of Always, Ask First, and Never action tiers for robust agent control.
  • Use Case: When starting a new AI agent project, use this Skill to ensure the agent's purpose, behavior, and constraints are clearly defined before implementation, preventing common pitfalls of vague or incomplete specifications.

Quick Start

Use the agent-spec-writer skill to start designing a new AI agent specification.

Frequently Asked Questions about agent-spec-writer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write an AI agent specification with clear behavioral boundaries?▼

To write an AI agent specification, you can use a structured six-phase process that guides you through defining a vision, PRD sections, and three-tier action boundaries to ensure clarity and safety.

What is the three-tier boundary system for AI agent guardrails?▼

The three-tier boundary system for agent guardrails categorizes actions into Always, Ask First, and Never tiers, enforcing explicit behavioral control and preventing unsafe or unintended operations.

Can I generate PRDs and specs for different frameworks like Claude and OpenCode?▼

Yes, you can generate specifications for multiple formats including SKILL.md, Claude agents, OpenCode agents, generic PRDs, and GitHub Spec Kits after completing the interactive design phases.

How do I define measurable success criteria when designing an AI agent?▼

Defining measurable success criteria involves following a structured validation phase during the agent design process, ensuring the specification includes verifiable metrics for the agent's performance.

When do I need a formal specification for my AI agent project?▼

You need a formal AI agent specification when starting a new project to ensure the agent's purpose, behavior, and constraints are clearly defined before implementation, preventing vague or incomplete requirements.

What's the best way to structure a product requirements document for AI development?▼

The best way to structure a PRD for AI development is through a phased approach covering orientation, vision, structure, guardrails, and validation, resulting in comprehensive and reliable agent specifications.

Related Skills

Explore More Software Engineering Skills→
Davie521Davie521
7

teach-impeccable

Scan a codebase and write design context to .impeccable.md.

Community
Advanced
xingwudaoxingwudao
116

build-strategy-spec

Convert strategy ideas into validated strategy_spec.yaml files using canonical recipes.

Community
Advanced
AesopScottAesopScott

window-map

Implement per-model context packing strategies and boundary rules for LLM integration.

Community
Advanced
whichguywhichguy

spike

Run disposable experiments to validate feasibility and compare approaches.

Community
Intermediate