skills.rest
gozonerd
gozonerdCOMMUNITY
Explore all skills by this author →
GitHub

Meta-Prompt Generation

Generate YAML meta-prompts and execution briefs for downstream AI agents.

Updated Apr 11, 2026
Search Tags:#validation#ai agents#yaml#guardrails#prompt generation#meta-prompt#instruction brief
One-click install
npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill meta-prompt-generation-gozonerd
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: Meta-Prompt Generation
Source: https://github.com/gozonerd/shadow-ai-assessment/tree/main/.claude/skills/dmis/task_type/TT-05_Meta_Prompt_Generation
Command: npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill meta-prompt-generation-gozonerd

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve? Downstream AI agents fail or produce inconsistent results when given vague instructions. This Skill produces precisely-specified meta-prompts and execution briefs in YAML so downstream agents can execute specialized tasks independently with clear quality guardrails. ## Core Features & Use Cases - Structured Brief Generation: Creates YAML instruction documents with task overview, input specification, output requirements, quality criteria, guardrails, and examples conforming to the TT-05 schema. - Embedded Guardrails: Enforces 9 accuracy rules (ACC-001 to ACC-009) and 6 anti-fabrication rules (AFR-001 to AFR-006) inside every generated prompt. - Output Validation: Includes a Python script that validates generated meta-prompts against required fields, task type, version, and guardrail presence. - Use Case: A pipeline coordinator needs an AI agent to synthesize domain knowledge on digital health. Use this Skill to generate a complete execution brief with citations requirements, accuracy constraints, and gold-standard examples. ## Quick Start Generate a TT-05 meta-prompt brief for a downstream agent that must summarize research evidence with source attribution and anti-fabrication guardrails.

Frequently Asked Questions about Meta-Prompt Generation

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

FAQPage Schema
How do I write a meta-prompt for a downstream AI agent?▼

Define the task overview, input specification, output requirements, quality criteria, guardrails, and examples in a YAML document. This Skill structures those sections so the downstream agent can execute without additional interpretation.

What format should AI agent execution briefs use?▼

Execution briefs use YAML conforming to the TT-05 schema with required fields: title, task_type, version, task_overview, input_specification, output_requirements, quality_criteria, and guardrails. The included validate_output.py script checks conformance.

How do I validate a generated meta-prompt YAML file?▼

Run scripts/validate_output.py with the file path as argument. It checks YAML parsing, required fields, task_type value TT-05, version containing v02, and the presence of anti-fabrication rules in guardrails.

What are anti-fabrication rules in AI prompts?▼

Anti-fabrication rules (AFR-001 to AFR-006) require every factual claim to have a traceable source, prohibit invented statistics or citations, and mandate flagging uncertain claims as [NEEDS VERIFICATION]. They are embedded in the guardrails section of each generated brief.

When should I not use a meta-prompt generation approach?▼

Avoid it for tasks the downstream agent cannot perform within its capabilities, or when over-specification would remove needed agent judgment. The Skill lists over-specification and unachievable tasks as anti-patterns.

Related Skills

Explore More Product & Management Skills→
ax-llmax-llm
2.9k

ax-rust-agent-context

Select context strategies for Rust Ax agents using axllm APIs and runtime profiles.

Official
Intermediate
m2ai-portfoliom2ai-portfolio
4

dispatch-handoff-brief

Generates structured delegation briefs with seven sections for autonomous AI agents.

Official
Advanced
genesis-agentsgenesis-agents
2

Verifier

Cross-check claims against cited sources and return verdicts with evidence.

Community
Advanced
HugoAlmeidaMoreiraHugoAlmeidaMoreira

outlines

Generate structured JSON outputs from natural language prompts using local models.

Community
Advanced