A ready-to-run Claude Code skill that pulls sentiment, topics, and action recommendations from TikTok, Instagram, and Twitter links using Apify under the hood.
This guide is reviewed for clarity, service accuracy, and AI-search readability. The next quarterly content review is tracked internally before unsupported metrics or client proof are added.
What This Skill Does
This is a Claude Code skill that accepts a list of social media URLs and returns structured intelligence on each one. For every TikTok, Instagram, or Twitter link you feed it, you get back the post's sentiment (positive, negative, or neutral), the topics it covers, a relevance score against your current work, and specific action recommendations with priority rankings and time estimates.
The heavy lifting runs through Apify. You supply the URLs, the skill handles the API calls, and the results land in a dated JSON file you can act on immediately.
Why It Matters for AI-Assisted Workflows
Manually reviewing competitor posts is slow and leaves nothing recorded. This skill turns that into a structured pipeline: drop in links after your morning scroll and get back a ranked action list in minutes.
Inside Claude Code it fits naturally into content and research workflows. You can chain it with a writing skill so Claude analyzes competitor posts and then drafts response content in the same session.
How to Install It
bashgit clone https://github.com/grandamenium/social-media-insights.git ~/.claude/skills/social-media-insights ls ~/.claude/skills/social-media-insights/
Setting It Up in Claude Code
Once the folder is in place, paste this prompt into Claude Code to configure and run a first test:
promptLoad the social-media-insights skill from ~/.claude/skills/social-media-insights/ and read the SKILL.md. Then create an input.json file in the current directory with two test URLs — one TikTok and one Instagram post from a competitor or topic I care about. Run the pipeline with that input and return the full output JSON.
Setting It Up in Codex
The setup in Codex is identical — the skill runs as a bash pipeline with no model-specific dependencies. Drop the skill folder into your Codex skills directory and use the same prompt above, replacing 'Claude Code' with 'Codex'.
Example Use Cases
Practical ways to fit this into a weekly workflow:
- Drop in 10 competitor reels every Monday and get a ranked list of topics gaining traction in your space
- Feed it posts from your own account to score which content themes are landing versus falling flat
- Pipe the recommendations output into a content brief skill to turn trend signals into draft ideas automatically
- Track a topic over time by saving dated JSON output files and comparing them week over week
- Use the priority_score field to cut noise — skip anything below 0.5, action anything above 0.8
Frequently Asked Questions
- Is this free to run? The skill itself is free and open source. Apify charges per API call — typical usage runs a few cents per post. A free Apify account covers testing.
- Does it work with YouTube or LinkedIn? The current version supports TikTok, Instagram, and Twitter/X only. YouTube and LinkedIn require different Apify actors not included in this skill.
- What happens if a post URL is deleted or private? The skill logs a failed status for that entry and continues with the rest of the batch. You will see an error note in the output JSON next to any URL that did not resolve.