prompt

Builds structured AI prompts through type classification, user interviews, and Anthropic best-practice techniques.

Updated Jun 8, 2026
One-click install
npx skills add https://github.com/shawn-sandy/agentics-kit --skill prompt-shawn-sandy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: prompt
Source: https://github.com/shawn-sandy/agentics-kit/tree/main/kit/plugins/plan-agent/skills/prompt
Command: npx skills add https://github.com/shawn-sandy/agentics-kit --skill prompt-shawn-sandy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Writing effective AI prompts requires knowing which prompting techniques apply to which kind of request, and most users either under-specify their intent or over-constrain the model. This Skill interviews you about your goal, classifies the prompt into one of five types, and assembles a copy-pasteable prompt grounded in Anthropic's published best practices. ## Core Features & Use Cases - Type classification and technique matrix: Classifies your need as system, task, creative, analytical, or proposal, then applies the matching set of techniques such as role assignment, XML structure, chain-of-thought scaffolding, and document grounding. - Guided interview: Asks 2-3 type-specific questions that extract the why behind your request, not just the what, so the resulting prompt carries real context and motivation. - Template-based drafting and saving: Fills a per-type template, recommends installed tools that might replace a custom prompt, and saves the result as a dated markdown file under a resolved prompts directory. - Use Case: Ask for help writing a system prompt for a customer support chatbot; the Skill confirms the type, asks about persona, tone, and boundaries, then delivers and saves a structured prompt with role, instructions, and guardrail sections. ## Quick Start Ask the assistant to help you write a prompt for your task, for example by describing the chatbot, refactoring job, or analysis you want the prompt to accomplish.

Frequently Asked Questions about prompt

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

FAQPage Schema
How do I write a better AI prompt for a specific task?▼

Describe your goal and let the Skill classify it into one of five prompt types, then answer 2-3 interview questions about your intent and constraints. It assembles a structured prompt using techniques like XML tags, role assignment, and output formatting, and delivers it in a copy-pasteable block.

What prompt types does this Skill support?▼

It supports five types: system prompts for assistants and agents, one-shot task instructions, creative writing prompts, analytical prompts for document analysis, and proposal prompts that carry grounded decisions. Each type selects a different technique matrix and template.

Can I skip the interview when generating a prompt?▼

Yes, callers that have already gathered answers can pass the --answers-gathered flag to bypass the interview phase entirely. The provided content is then fed directly into the structuring and drafting phases.

Where are generated prompts saved?▼

Prompts are saved as markdown files in a directory resolved from the promptsDirectory setting or the git root's docs/prompts folder. Filenames follow a type-intent-slug-date pattern, and you can override everything with the --out flag.

What happens if a saved prompt file already exists?▼

For most types the Skill appends a numeric suffix to keep filenames unique. For proposal prompts it rewrites in place, but first checks a recorded sha256 hash and asks for confirmation if the file was hand-edited since the last generation.