skills.rest
thorsenk
thorsenkCOMMUNITY
Explore all skills by this author →
GitHub

prompt-maestro

Classify, audit, and harden prompts with a visual HTML scorecard report.

Updated Jun 29, 2026
Search Tags:#prompt engineering#scorecard#prompt evaluation#quality review#html report#prompt hardening#prompt audit
One-click install
npx skills add https://github.com/thorsenk/skills --skill prompt-maestro-thorsenk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: prompt-maestro
Source: https://github.com/thorsenk/skills/tree/main/skills/prompt-maestro
Command: npx skills add https://github.com/thorsenk/skills --skill prompt-maestro-thorsenk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Teams often run prompts that are ambiguous, over-scoped, or missing safety boundaries, leading to wasted runs and risky outputs. This Skill reviews any prompt before execution, classifies it, scores it against a structured rubric, and returns exact fixes plus a visual HTML report. ## Core Features & Use Cases - Prompt Classification: Assigns primary type, context dependency, lifecycle, target executor, and risk level using a defined taxonomy. - Rubric-Based Audit: Scores ten dimensions including objective clarity, scope boundaries, autonomy, output contract, and risk handling, with score caps for blocking issues. - Visual HTML Scorecard: Produces a standalone long-scroll report with score rings, severity chips, findings, copy-ready edits, and a revised prompt. - Use Case: Before handing a workspace-bound research prompt to an agent, run this Skill to catch missing search boundaries and evidence requirements, then apply the suggested replacement language. ## Quick Start Use prompt-maestro to review this prompt and produce an HTML scorecard report before anyone runs it.

Frequently Asked Questions about prompt-maestro

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

FAQPage Schema
How do I audit a prompt before running it?▼

Paste the prompt and ask for a review. The Skill classifies the prompt type, scores it against a ten-dimension rubric covering clarity, scope, autonomy, and risk handling, then returns findings with exact replacement language and an HTML scorecard report.

What types of prompts can be evaluated?▼

The taxonomy covers audit, implementation, research, extraction, classification, planning, creative, analysis, agent/skill, evaluation, and automation prompts. Each is also classified by context dependency, lifecycle, and risk level from low to critical.

Does the prompt review execute the target prompt?▼

No. The Skill acts only as a reviewer and report producer and never runs the target prompt during evaluation. Execution happens only if the user explicitly asks for it after the review is complete.

How is the prompt quality score calculated?▼

Scores range from 0 to 100 across rubric dimensions like objective clarity, output contract, and risk handling. Caps apply: 59 if the prompt could solve the wrong problem, 74 if high-risk actions lack boundaries, and 89 if the output contract is incomplete.

When should I not use a prompt review workflow?▼

Skip it when you want the prompt executed immediately rather than reviewed, when you need general writing help with no reusable prompt, or when the task is visual storyboard prompt orchestration, which belongs to a dedicated storyboard package.

Related Skills

Explore More Product & Management Skills→
hujianbesthujianbest
1

hf-bug-patterns

Identifies recurring mistakes and drafts a bug pattern candidate for human confirmation.

Community
Intermediate
Uomi-networkUomi-network

wasp-call

Invoke a deployed UOMI agent on-chain by NFT ID and retrieve output.

Official
Advanced
RefoundAIRefoundAI
1.2k

conducting-interviews

Train teams to conduct consistent product interviews with structured rubrics.

Official
Advanced
lshtramlshtram

meta-prompting

Design optimal prompts and task breakdowns for complex AI reasoning.

Community
Intermediate