/analyze-prompt

Analyze a prompt or system prompt for improvement opportunities using state-of-the-art prompt engineering principles from 2026.

Analysis Framework

1. Gather the Prompt

First, obtain the prompt to analyze:

• If the user provides a file path, read it with the Read tool
• If the user pastes it directly, use that
• If no prompt provided, use AskUserQuestion to request it

2. Structural Analysis

Evaluate the prompt structure using the TCRTE Framework (Task, Context, Role, Tone, Examples):

Component	Present?	Quality (1-5)	Notes
Task	Clear directive of what to do
Context	Background information and constraints
Role	Who the model should act as
Tone	Style and voice guidance
Examples	Few-shot demonstrations

3. Clarity Assessment

Check for common clarity issues:

• Ambiguity: Vague instructions that could be interpreted multiple ways
• Missing constraints: Undefined scope, length, or format
• Conflicting instructions: Contradictory guidance
• Assumed context: References to information not provided
• Implicit expectations: Behaviors expected but not stated

4. Claude 4.x Compatibility (2026)

Claude 4.x models follow instructions precisely. Check for:

• Explicit action requests: "Create X" vs "Can you suggest X?"
• Specificity: Exact format, length, and style specified
• Context motivation: Explains WHY instructions matter
• Example alignment: Examples match desired output exactly
• Negative framing: "Don't do X" should be "Do Y instead"

5. Context Window Efficiency

Evaluate token efficiency:

• Redundancy: Repeated information
• Verbosity: Overly wordy sections that could be condensed
• Dead weight: Information that doesn't contribute to task success
• Missing high-signal content: Important context that's absent

6. Advanced Pattern Check

Look for opportunities to apply:

• Structured output: Would JSON/XML output improve reliability?
• Chain-of-thought: Would step-by-step reasoning help?
• Few-shot examples: Would 3-5 examples dramatically improve output?
• Prompt scaffolding: Does it need defensive guardrails?
• XML delineation: Would structured sections improve parsing?

7. Generate Report

Produce a structured analysis report:

## Prompt Analysis Report

### Overall Score: X/10

### Strengths
- [List what the prompt does well]

### Critical Issues
1. **[Issue name]**: [Description and impact]
   - **Fix**: [Specific recommendation]

### Optimization Opportunities
1. **[Opportunity]**: [How it would improve results]

### Recommended Rewrite
[If significant changes needed, provide optimized version]

### Quick Wins
- [Small changes with high impact]

### Token Efficiency
- Current estimated tokens: ~X
- Potential reduction: ~Y tokens (Z%)
- Key areas for compression: [List]

Common Patterns to Flag

Red Flags

• "Do whatever you think is best" (too vague)
• "Be creative" without constraints (undefined scope)
• Long lists of "don't do" instructions (should be positive framing)
• Examples that don't match desired output format
• Missing success criteria

Green Flags

• Clear task statement in first sentence
• Explicit format specification
• Motivational context for instructions
• Diverse, high-quality examples
• Specific success criteria

Example Analysis

Poor Prompt:

Help me write better code

Analysis:

• Task: Vague ("better" is undefined)
• Context: None
• Role: None
• Tone: None
• Examples: None
• Score: 2/10

Optimized:

Review this Python function and improve its performance, readability, and error handling.

Focus on:
1. Time complexity optimization
2. PEP 8 style compliance
3. Proper exception handling with specific exception types
4. Clear variable names and docstrings

Return the improved code with inline comments explaining each change.

<code>
[code here]
</code>

Score: 8/10 - Clear task, specific focus areas, defined output format