Review Suno Song

Mode 2: Review Direct Text

When pasting prompt + lyrics directly:

/review-song

The skill will:

1. Prompt you to paste your structured prompt
2. Prompt you to paste your lyrics
3. Extract genre/mood context
4. Launch quality-reviewer sub-agent
5. Present structured feedback

What Gets Evaluated

Prompt Quality (Structured Sections)

Structure:

• Proper colon-and-quotes format
• No blank lines between sections
• Required sections present (genre, vocal, instrumentation, production, mood)

Specificity:

• Concrete descriptors vs. vague abstractions
• Technical vocabulary appropriate to genre
• Clear production techniques described

Copyright Safety:

• No artist names
• No band names
• No album titles
• No song titles
• Style descriptions focus on characteristics

Genre Alignment:

• Descriptors match genre conventions
• No contradictory elements
• Appropriate technical vocabulary

Lyric Quality (Comprehensive Assessment)

AI-Slop Detection:

• Technology clichés: "neon lights", "digital", "echoes in the void"
• Abstract vagueness: "whispers in the dark", "fragments of", "fading memories"
• Generic imagery without concrete context

Cliché Detection:

• Romantic clichés: "heart on my sleeve", "falling for you", "love at first sight"
• General song clichés: "time will heal", "reach for the stars", "follow your dreams"
• Genre-specific clichés: Country (trucks/beer), Pop (dancing all night), Rock (breaking chains)
• Lazy rhyming with cliché phrases

Poor Lyric Quality:

• Awkward or clunky phrasing that doesn't flow
• Grammatical issues (unless intentional for style)
• Nonsensical or confusing imagery
• Mixed metaphors that contradict
• Lines that are too wordy or verbose
• Unintentionally funny or cringe-worthy lines
• Excessive filler words ("yeah yeah yeah" without purpose)
• Trying too hard to be clever/poetic and failing
• Inconsistent voice or jarring tone shifts

Specificity vs. Abstractions:

• Concrete nouns, specific numbers, physical details
• Sensory details vs. vague generalities
• "Show don't tell" principle

Metaphor Consistency:

• Central metaphor maintained throughout
• No contradictory imagery
• Coherent metaphor system

Syllable Patterns:

• Consistency within sections
• Singability without awkward rushing
• Natural emphasis patterns

Rhyme Scheme and Quality:

• Pattern identification (AABB, ABAB, ABCB, etc.)
• Rhyme quality (exact, slant, forced)
• Genre appropriateness
• Avoidance of over-reliance on easy rhymes

Style-Lyric Consistency:

• Content matches genre expectations
• Tone alignment (playful pop vs. serious ballad)
• Language complexity appropriate for style
• Subject matter fits genre conventions

Gender-Pronoun Consistency:

• POV clarity for vocalist gender
• Narrative context for pronoun usage
• Check for confusing or contradictory pronouns

General Taste and Quality:

• Catchiness and memorability
• Flow and singability
• Emotional resonance and authenticity
• Hook strength
• Professional polish vs. amateur feel

Output Format

Receive structured feedback categorized by severity:

**Prompt Quality: X/10**
- Structure: [✓/⚠️/✗] [comment]
- Specificity: [✓/⚠️/✗] [comment]
- Copyright: [✓/✗] [comment]
- Genre alignment: [✓/⚠️/✗] [comment]

**Lyric Quality: X/10**
- AI-slop: [count] instances - [specific examples with line numbers]
- Clichés: [count] instances - [specific examples with line numbers]
- Poor quality lines: [count] instances - [specific examples with line numbers and reasons]
- Specificity: [✓/⚠️/✗] [comment]
- Metaphor consistency: [✓/⚠️/✗] [comment]
- Syllable patterns: [✓/⚠️/✗] [comment]
- Rhyme scheme: [✓/⚠️/✗] [pattern and quality assessment]
- Style-lyric fit: [✓/⚠️/✗] [genre expectations match]
- Gender-pronoun consistency: [✓/⚠️/✗] [POV clarity]
- General taste: [X/10] [overall quality assessment]

**Recommendations (by severity):**

CRITICAL (must fix):
1. [Specific issue with line numbers and reasoning]

SUGGESTED (strong recommendations):
1. [Specific improvement with suggested replacement]

OPTIONAL (nice-to-have):
1. [Refinement suggestion with reasoning]

How It Works

Internal Process

1. Extract Content:

   • If file path provided: Read file, parse YAML frontmatter, extract prompt sections and lyrics
   • If direct text: Prompt user to paste content

2. Identify Context:

   • Extract genre from prompt
   • Extract mood from prompt
   • Extract vocal style from prompt

2.5. Ask Genre-Specific Refinement Questions (NEW):

Use AskUserQuestion tool to collect evaluation preferences from user.

Question 1: Specificity Preference