$npx -y skills add ScrapeCreators/social-media-research-skills --skill comment-miningUse when the user wants to mine comments and replies for audience reactions, customer language, questions, objections, complaints, product ideas, buying intent, sentiment, or voice-of-customer insights from public social posts and videos.
| 1 | # Comment Mining |
| 2 | |
| 3 | ## Overview |
| 4 | |
| 5 | Mine public comments for what people actually ask, complain about, want, misunderstand, or repeat. The output should help with product research, content ideas, copywriting, objection handling, and audience understanding. |
| 6 | |
| 7 | ## When to Use |
| 8 | |
| 9 | Use this skill when the user asks to: |
| 10 | |
| 11 | - analyze comments on a TikTok, YouTube video, Instagram Reel, Facebook post, Reddit post, or Rumble video |
| 12 | - find audience questions, objections, complaints, or buying intent |
| 13 | - extract voice-of-customer language |
| 14 | - find content ideas from comments |
| 15 | - understand sentiment around a post, creator, product, or topic |
| 16 | |
| 17 | ## Comment Sources |
| 18 | |
| 19 | | Platform | Endpoint | |
| 20 | |---|---| |
| 21 | | TikTok comments | `/v1/tiktok/video/comments` | |
| 22 | | TikTok replies | `/v1/tiktok/video/comment/replies` | |
| 23 | | YouTube comments | `/v1/youtube/video/comments` | |
| 24 | | YouTube replies | `/v1/youtube/video/comment/replies` | |
| 25 | | Instagram comments | `/v2/instagram/post/comments` | |
| 26 | | Facebook comments | `/v1/facebook/post/comments` | |
| 27 | | Facebook replies | `/v1/facebook/post/comment/replies` | |
| 28 | | Reddit comments | `/v1/reddit/post/comments` | |
| 29 | | Rumble comments | `/v1/rumble/video/comments` | |
| 30 | |
| 31 | ## Workflow |
| 32 | |
| 33 | 1. **Fetch comments** |
| 34 | - Use the post/video URL whenever possible. |
| 35 | - Paginate when the endpoint supports it and the user wants depth. |
| 36 | - Preserve comment text, author if public, like/upvote count, timestamp, and source URL. |
| 37 | |
| 38 | 2. **Clean lightly** |
| 39 | - Remove obvious spam/duplicates. |
| 40 | - Keep slang, misspellings, and emotional wording if it is useful customer language. |
| 41 | - Do not over-normalize exact quotes. |
| 42 | |
| 43 | 3. **Classify each useful comment** |
| 44 | Use these buckets: |
| 45 | - questions |
| 46 | - objections |
| 47 | - complaints/pain points |
| 48 | - praise |
| 49 | - confusion |
| 50 | - requests/feature ideas |
| 51 | - buying intent |
| 52 | - controversy/debate |
| 53 | - jokes/memes/culture signals |
| 54 | |
| 55 | 4. **Cluster themes** |
| 56 | - Group similar comments. |
| 57 | - Score themes by frequency and intensity. |
| 58 | - Highlight exact quotes for each theme. |
| 59 | |
| 60 | 5. **Turn insights into actions** |
| 61 | Depending on the user's goal, produce: |
| 62 | - content ideas |
| 63 | - FAQ ideas |
| 64 | - landing page copy angles |
| 65 | - product ideas |
| 66 | - objection-handling bullets |
| 67 | - sales/support notes |
| 68 | |
| 69 | ## Output Format |
| 70 | |
| 71 | ```markdown |
| 72 | # Comment Mining Report |
| 73 | |
| 74 | ## Summary |
| 75 | - Source(s): {urls} |
| 76 | - Comments analyzed: {count} |
| 77 | - Confidence: High/Medium/Low |
| 78 | |
| 79 | ## Top Themes |
| 80 | | Theme | Type | Frequency | Intensity | Representative quote | |
| 81 | |---|---|---:|---|---| |
| 82 | |
| 83 | ## Audience Questions |
| 84 | - "..." |
| 85 | |
| 86 | ## Objections and Concerns |
| 87 | - **Objection:** ... |
| 88 | - Evidence: "..." |
| 89 | - Response angle: ... |
| 90 | |
| 91 | ## Buying Intent / Demand Signals |
| 92 | - "..." |
| 93 | |
| 94 | ## Exact Language to Reuse |
| 95 | - "..." |
| 96 | - "..." |
| 97 | |
| 98 | ## Content Ideas From Comments |
| 99 | 1. ... |
| 100 | 2. ... |
| 101 | ``` |
| 102 | |
| 103 | ## Quality Guardrails |
| 104 | |
| 105 | - Label sample size and confidence. |
| 106 | - Separate one loud comment from a repeated pattern. |
| 107 | - Preserve exact quotes for useful language. |
| 108 | - Avoid claiming broad market sentiment from one post's comments. |
| 109 | - Call out moderation/platform bias when relevant. |
| 110 | |
| 111 | ## Common Pitfalls |
| 112 | |
| 113 | - Do not flatten comments into generic sentiment. The value is in questions, objections, and exact wording. |
| 114 | - Do not include personally identifying details unless they are already public and necessary. |
| 115 | - Do not treat bot/spam comments as audience signal. |
| 116 | - Do not skip Reddit post context. For Reddit, read both the original post and comments. |