meta-prompt

Rewrites rough prompts into scoped, project-aware prompts for AI coding assistants.

Updated Jun 11, 2026
One-click install
npx skills add https://github.com/Marcel-Carrillo/ProyectosIA --skill meta-prompt-marcel-carrillo
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: meta-prompt
Source: https://github.com/Marcel-Carrillo/ProyectosIA/tree/main/ai-specs/skills/meta-prompt
Command: npx skills add https://github.com/Marcel-Carrillo/ProyectosIA --skill meta-prompt-marcel-carrillo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Vague or underspecified prompts cause AI coding assistants to produce off-target results, touch unrelated files, or ignore project constraints. This Skill rewrites rough prompts into precise, scoped, actionable prompts aligned with the project's OpenSpec spec-driven development workflow. ## Core Features & Use Cases - Structured Prompt Rewriting: Transforms raw prompts using a defined template covering role, objective, scope, constraints, expected output, verification, and guardrails. - Prompt Type Templates: Provides dedicated structures for OpenSpec change prompts, implementation prompts, debugging prompts, code review prompts, documentation prompts, and Git/commit prompts. - Selective Context Loading: Injects only the relevant standards files (backend, frontend, product, documentation, or Git) instead of telling the assistant to read the whole repository. - Use Case: A developer pastes "add product catalog management" and receives a complete OpenSpec change prompt with business rules, acceptance criteria, non-goals, and explicit instructions not to implement code yet. ## Quick Start Ask the assistant to rewrite your rough prompt, for example: improve this prompt for implementing product catalog management in Claude Code.

Frequently Asked Questions about meta-prompt

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

FAQPage Schema
How do I write a better prompt for Claude Code or Cursor?▼

Provide your rough prompt and this Skill rewrites it with a defined structure covering role, objective, scope, constraints, expected output, and verification. It adds relevant project context and guardrails while preserving your original intent.

How to create a prompt for an OpenSpec change proposal?▼

Use the OpenSpec change prompt type, which includes change name, business goal, actors, functional requirements, acceptance criteria, API and data model impact, and non-goals. It explicitly instructs the assistant not to implement code or create commits.

Does this Skill execute the rewritten prompt?▼

No, it only rewrites the prompt and returns it as text. It does not edit files, run commands, create specs, implement code, or create commits.

Can the rewritten prompts include external tools like Jira or Sentry?▼

Only when explicitly requested. GitHub is included when branches, commits, or PRs are relevant, and Context7 only when available and official documentation is needed. Jira and Sentry are excluded by default.

Why does the Skill limit which documentation files the prompt references?▼

It enforces selective context loading so the assistant reads only task-relevant standards files, such as backend-standards.md for backend work. This avoids wasting context on the whole repository and keeps responses focused.