senior-prompt-engineer
World-class prompt engineering skill for LLM optimization, prompt patterns, structured outputs, and AI product development. Expertise in Claude, GPT-4, prompt design patterns, few-shot learning, chain-of-thought, and AI evaluation. Includes RAG optimization, agent design, and LLM system architecture. Use when building AI products, optimizing LLM performance, designing agentic systems, or implementing advanced prompting techniques.
Install with Tessl CLI
npx tessl i github:sc30gsw/claude-code-customes --skill senior-prompt-engineer50
Does it follow best practices?
If you maintain this skill, you can automatically optimize it using the tessl CLI to improve its score:
npx tessl skill review --optimize ./path/to/skillValidation for skill structure
Discovery
82%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
The description has strong trigger term coverage and completeness with an explicit 'Use when' clause. However, it relies heavily on buzzwords ('world-class', 'expertise') rather than concrete actions, and its broad scope covering multiple AI domains could create conflicts with other skills. The description would benefit from focusing on specific actions rather than claiming expertise.
Suggestions
Replace buzzword claims ('world-class', 'expertise in') with concrete actions (e.g., 'Writes and refines prompts', 'Debugs LLM outputs', 'Designs prompt templates')
Consider narrowing scope or clarifying boundaries - currently covers prompts, RAG, agents, and evaluation which could each be separate skills
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (prompt engineering, LLM optimization) and lists several techniques (few-shot learning, chain-of-thought, RAG), but uses buzzword-heavy language like 'world-class' and 'expertise' rather than concrete actions. Does not specify what actions are performed (e.g., 'writes prompts', 'debugs outputs'). | 2 / 3 |
Completeness | Explicitly answers both what (prompt engineering, LLM optimization, prompt patterns, structured outputs) and when ('Use when building AI products, optimizing LLM performance, designing agentic systems, or implementing advanced prompting techniques'). Has clear 'Use when...' clause. | 3 / 3 |
Trigger Term Quality | Good coverage of natural terms users would say: 'prompt engineering', 'Claude', 'GPT-4', 'few-shot', 'chain-of-thought', 'RAG', 'agent design', 'AI products'. These are terms users would naturally use when seeking help with LLM work. | 3 / 3 |
Distinctiveness Conflict Risk | While prompt engineering is a specific domain, terms like 'AI product development' and 'LLM system architecture' are broad and could overlap with general coding or architecture skills. The scope is quite wide, covering prompts, RAG, agents, and evaluation. | 2 / 3 |
Total | 10 / 12 Passed |
Implementation
7%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is fundamentally broken - it claims to be about prompt engineering but contains zero actual prompt engineering content. Instead, it's filled with generic software engineering platitudes, fake script references, and verbose lists of concepts Claude already knows. The skill provides no actionable guidance on writing prompts, structuring LLM interactions, or implementing any prompting techniques.
Suggestions
Replace the entire content with actual prompt engineering patterns - show concrete examples of few-shot prompting, chain-of-thought, structured output schemas with real, executable examples
Remove all generic software engineering content (TDD, code reviews, team leadership) that Claude already knows and that isn't specific to prompt engineering
Add concrete, copy-paste-ready prompt templates with input/output examples for common tasks like classification, extraction, and generation
Include actual validation workflows for prompt testing - how to evaluate prompt quality, iterate on failures, and measure improvements with specific metrics and commands
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose with extensive padding. Lists generic concepts Claude already knows (test-driven development, code reviews, authentication), includes irrelevant tech stack enumeration, and provides vague bullet points instead of actionable content. The 'Senior-Level Responsibilities' section is entirely unnecessary filler. | 1 / 3 |
Actionability | Despite claiming to be a prompt engineering skill, contains zero actual prompt engineering guidance. The code examples are fake placeholder scripts that don't exist. No concrete prompting patterns, no executable examples, no specific techniques - just abstract descriptions like 'Advanced patterns and best practices' without any actual patterns. | 1 / 3 |
Workflow Clarity | No clear workflows for any prompt engineering task. The 'Production Patterns' section lists vague concepts without steps. No validation checkpoints, no sequenced processes, no feedback loops. A user would have no idea how to actually engineer a prompt after reading this. | 1 / 3 |
Progressive Disclosure | References external files (references/prompt_engineering_patterns.md, etc.) which is appropriate structure, but the main file itself is a monolithic wall of generic content. The references are one level deep, but the signaling is weak and the overview content is bloated rather than concise. | 2 / 3 |
Total | 5 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.