Content
17%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is essentially a copy of a prompt template library that teaches Claude things it already knows — prompt engineering techniques, role patterns, and task-specific prompt structures. It consumes significant context window space without providing novel, actionable guidance. The content would be more appropriate as a human-facing reference document than as a Claude skill file.
Suggestions
Fundamentally reconsider the skill's purpose: Claude already knows how to write prompts. Focus instead on project-specific prompt conventions, preferred formats, or organization-specific patterns that Claude wouldn't know.
If the skill is meant to help Claude serve prompt templates to users, add a clear workflow: how to identify the user's need, select the right template category, customize placeholders, and present the result.
Split the monolithic content into separate category files (role-prompts.md, task-prompts.md, etc.) and make SKILL.md a concise index with one-line descriptions and links.
Remove prompt engineering technique explanations (CoT, few-shot, persona pattern, structured output) entirely — Claude knows these deeply and explaining them wastes tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | This skill extensively explains concepts Claude already knows deeply — prompt engineering techniques, role-based prompting, chain-of-thought, few-shot learning, etc. Claude is an LLM that inherently understands these patterns. Nearly every token here is teaching Claude things it already knows, making this extremely wasteful of context window. | 1 / 5 |
Actionability | The prompts are templates with placeholders like [CONCEPT], [FRAMEWORK], [SOURCE], [TARGET] but provide no concrete executable guidance for Claude itself. These are prompt templates for humans to use with LLMs, not actionable instructions for Claude to follow when performing a task. The skill tells Claude to serve as a prompt library but doesn't give concrete steps for how to select, customize, or deliver prompts. | 2 / 5 |
Workflow Clarity | There is no clear workflow for how Claude should use this library. The 'When to Use This Skill' section lists triggers but doesn't define a process — e.g., how to select the right prompt category, how to customize templates for the user's specific need, or how to validate the output. The checklist at the end is aimed at humans, not at Claude's workflow. | 2 / 5 |
Progressive Disclosure | This is a monolithic wall of prompt templates (~250 lines) all inlined in a single file with no references to supporting files. The content would benefit enormously from being split into category-specific files (e.g., role-prompts.md, task-prompts.md, analysis-prompts.md) with SKILL.md serving as an overview/index. The external resources linked are general web URLs, not bundle files. | 2 / 5 |
Total | 7 / 20 Passed |