Agent skill

Feedback Integration

Process viewer comments and feedback to improve content

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/feedback-integration

SKILL.md

Feedback Integration Skill

Listening to Our Listeners

This skill processes viewer comments and direct feedback to understand impact and improve future sessions.


Purpose

Extract qualitative insights from viewer feedback and integrate into the knowledge base.


Command

bash
/learn-comments session-name

Input Sources

Source Type Value
YouTube Comments Public Unfiltered reactions
Email Feedback Private Detailed testimonials
Social Mentions Public Organic shares
Direct Messages Private Personal experiences

Comment Analysis Process

1. Collection

Gather comments from YouTube Studio or API:

yaml
comments:
  - author: "Seeker123"
    text: "This helped me sleep for the first time in weeks..."
    likes: 42
    timestamp: "2025-01-10T14:32:00Z"
  - author: "MeditatorPro"
    text: "The Eden imagery was so vivid, I felt like I was there"
    likes: 28
    timestamp: "2025-01-11T09:15:00Z"

2. Categorization

Classify each comment:

Category Indicators
Impact "helped me", "changed my", "finally"
Experience "felt like", "saw", "experienced"
Technical "audio", "sound", "voice", "quality"
Request "could you", "please make", "I wish"
Question "how do I", "what is", "when should"
Criticism "didn't like", "too long", "confusing"

3. Sentiment Analysis

Score overall sentiment:

Sentiment Score Range Meaning
Very Positive 0.8-1.0 Strong resonance
Positive 0.5-0.8 Good reception
Neutral 0.3-0.5 Mixed signals
Negative 0.0-0.3 Issues to address

4. Theme Extraction

Identify recurring themes:

yaml
themes:
  - theme: "sleep_improvement"
    frequency: 12
    sentiment: 0.9
    sample_quotes:
      - "Finally slept through the night"
      - "Best sleep aid I've found"

  - theme: "vivid_imagery"
    frequency: 8
    sentiment: 0.85
    sample_quotes:
      - "The garden felt so real"
      - "I could actually see the light"

5. Insight Generation

Convert themes to actionable insights:

yaml
insight:
  finding: "Eden garden imagery creates exceptionally vivid experiences"
  evidence: "8 comments specifically mentioned vivid imagery in Eden session"
  action: "Use garden/nature settings for healing sessions"
  confidence: medium
  source: "comments"

Feedback Categories

Impact Testimonials

Most valuable for understanding effectiveness:

"I've struggled with anxiety for years, and this session gave me the first deep relaxation I've felt in months."

Extract:

  • Outcome achieved: anxiety relief
  • Session type: healing
  • Impact level: significant
  • Time element: long-term struggle resolved

Experience Reports

Valuable for refining content:

"When you described the waterfall, I could actually feel the cool mist on my skin."

Extract:

  • Effective element: waterfall imagery
  • Sensory channel: kinesthetic (touch)
  • Immersion level: high

Technical Feedback

Valuable for production quality:

"The voice was perfect but the binaural felt a bit too loud in my right ear."

Extract:

  • Voice quality: positive
  • Binaural mixing: needs review
  • Specific issue: channel balance

Requests

Valuable for content planning:

"I'd love a session focused on confidence before presentations."

Extract:

  • Requested outcome: confidence
  • Context: professional/presentations
  • Priority: add to topic queue

Negative Feedback Handling

Constructive Criticism

Address in future sessions:

Criticism Response
"Too slow at the start" Review induction pacing
"Voice felt monotonous" Add prosody variation
"Couldn't visualize" Enhance sensory descriptions
"Ending was abrupt" Extend emergence section

Non-Constructive

Note but don't over-weight:

Comment Response
"This is fake" Acknowledge different expectations
"Hypnosis is dangerous" Note misconception, don't change approach
Spam/trolling Ignore

Knowledge Base Updates

Store insights in knowledge/lessons_learned.yaml:

yaml
- id: "FEEDBACK-2025-001"
  category: "feedback"
  finding: "Kinesthetic imagery (touch sensations) increases immersion"
  evidence: "Multiple comments mention feeling physical sensations"
  action: "Include tactile descriptions in every visualization"
  confidence: medium
  source: "viewer_comments"
  sessions_referenced:
    - "eden-garden-pathworking"
    - "healing-waterfall-journey"
  date_discovered: "2025-01-15"

Feedback Loop Timing

Period Action
24-48 hours Initial comment surge
7 days First round analysis
30 days Comprehensive review
90 days Long-term impact assessment

Output

After analysis:

  1. Feedback Summary: Categorized comments with sentiment
  2. Theme Report: Recurring patterns with evidence
  3. Insight Updates: New entries for knowledge base
  4. Action Items: Specific improvements for next sessions
  5. Request Queue: Topics requested by viewers

Integration with Analytics

Combine feedback with metrics for complete picture:

Metric Says Feedback Says Conclusion
High retention "Loved every minute" True success
Low retention "Got distracted" Content issue
High retention "Fell asleep" Works for sleep (intended?)
Low likes "Too different for me" Niche content, not failure

Related Resources

  • Skill: tier4-growth/analytics-learning/ (quantitative)
  • Knowledge: knowledge/lessons_learned.yaml

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results