I Made a Longbridge Influencer Public Activity Research Skill

Recently, I created a personal-use Skill: Longbridge Influencer / Blogger Public Activity Research Assistant.

The problem it solves is simple: In the community, people often ask, "What is this blogger's trading logic?" "What are the specific prices for the stock acquisition/exercise/sell put they mentioned?" "Is this an official account?" Looking at just one or two posts can easily lead to taking things out of context. So, I've automated this process into a verifiable Skill.

What It Can Do

Given a Longbridge profile URL or member ID, it scrapes the blogger's public original posts and generates several types of files:

• Raw activity jsonl: Preserves paginated original text for easy review
  
• Tagged activities: Labels each market-related post, e.g., options, ETF, portfolio, macroeconomics, risk control, position
  
• Operation activities: Specifically filters out operational expressions like buy, sell, increase position, decrease position, exercise, stock acquisition, sell put
  
• evidence.tsv: Organizes main posts and public comments into an evidence table, including author, whether it's the OP, statement type, and original link
  
• trade-events.tsv: Extracts structured trading events only from the blogger's main posts and their own replies, avoiding misinterpreting others' questions as the blogger's trades
  

The core principle is: Only organizes publicly available information, does not equate public statements with actual transactions, and does not constitute investment advice.

Example 1: Analyzing Freedom Road's DRAM Sell Put

Previously, looking at Freedom Road's posts about DRAM, just reading the main post might make you think he only said, "Exercising means all-in on DRAM, about 50,000 shares."

But after scraping the comments together, you can see the OP himself added more crucial information in the replies:

• It's the combined total of two accounts
  
• Executed at three price points: 68, 66, and 64
  
• Someone asked in the context, "50,000 shares? That's selling 500 lots of puts?"
  
• Later, the OP also said, "This can only wait for exercise," and "I didn't buy the dip, I'm preparing to exercise."
  

The Skill breaks these down into evidence and trade-events. This way, you don't just imagine based on a single "all-in" statement, nor do you directly treat other users' comments as the OP's transactions.

Example 2: Researching LongbridgeAI and Option Master

I also used it to look at the LongbridgeAI and Option Master accounts.

The scraping results were roughly:

• LongbridgeAI: 31 posts, 450 comments
  
• Option Master: 98 posts, 673 comments
  

The conclusion is clear: LongbridgeAI is more like an official growth account for AI/CLI/Skill, focusing on tool entry points, workflows, events, and user cases. Option Master is more like an options education and product scenario account, focusing on P/C, OI, IV, triple witching days, pre-market options, portfolio options, and options sentiment radar.

There's another detail here: Option Master's articles often contain many "options case studies" and "market movement interpretations." The Skill extracts trade-events as clues, but the report won't present them as Option Master's own real trades. This is very important.

Where I Think This Skill Is Useful

It's not for "copy-trading influencers," but for turning public content into research material that can be examined:

• First, see what the blogger is consistently focused on long-term
  
• Then, look at their tool preferences: common stock, ETF, options, sell put, covered call, cash management
  
• Then, check if key comments contain supplementary position, price, strike price, expiration date details
  
• Finally, break down the quantifiable parts into rules, rather than directly copying opinions
  

For example, it can be summarized as:

• Selling puts isn't necessarily bearish; you need to consider "willingness to acquire stock/cash collateral/buying stock at a low price"
  
• The main post states the view, but the comments often contain price, lot size, and position details
  
• Options case studies in official educational accounts cannot be treated as personal trades
  
• Medium-to-low confidence trade-events must be cross-checked with the original evidence text
  

Data Boundaries

This Skill only processes publicly scrapable content:

• Does not cover private, deleted, or content invisible without login
  
• Does not guarantee 100% completeness of comment pagination
  
• If images only contain screenshots without OCR, they won't be automatically recognized
  
• Automatic classification is only an aid; key conclusions must be verified by reviewing the original links
  

If you also research community bloggers or track someone who consistently shares trading frameworks, this type of tool should save you a lot of time.

It's still being refined for personal use. If LongbridgeAI or the official team is interested in this type of community public content research Skill, I can continue to organize it into an installable zip/repo.