Most LinkedIn optimization advice is useless because it's vague. "Write a compelling headline." "Tell your story." People read that, nod, and then stare at their profile with no idea what to actually type. The advice lacks specificity, and specificity is the entire point. A recruiter doesn't think in vibes. They type keywords into a search filter and look at what comes back.
I built a LinkedIn score tool that takes your data export and scores it, section by section, against the mechanics of how recruiters actually find candidates. It tells you the exact problem with your headline, writes you a new one, and gives you the steps to change it. The whole report is an action plan you can execute in one sitting.
Where the idea came from
Credit goes to Nicole Barra, a recruiter who published a detailed guide on LinkedIn profile optimization for international job seekers. She walked through every profile section explaining what makes it work from the recruiter's side of the tool. Things like: recruiters filter by tech stack keywords in your headline, not your company name. An expired certification signals you stopped investing in yourself, which is worse than having no certification at all. Soft skills listed in your skills section take up space without matching any recruiter search filter.
Her guide was specific enough that I could codify it. Every insight mapped to a clear rule: if the headline contains only "Job Title at Company," it fails keyword search. If the About section is under 50 words, it provides no signal. If recommendations only come from same-level peers, there's no upward validation. These are binary checks against structured data, and LinkedIn conveniently lets you export your entire profile as CSV files.
How it works
You export your data from LinkedIn (Settings, Data Privacy, Download your data), unzip the archive, and point the tool at the folder. It reads Profile.csv, Positions.csv, Skills.csv, Recommendations Received.csv, and everything else in the export. Each section gets a weighted score based on its impact on recruiter visibility. Headline carries 15% of the total weight. Experience descriptions carry 12%. Skills carry 12%. The final output is a 0-100 score with a fix priority list ordered by impact.
The part that makes it more than a scorecard is the action plan. For every section scoring below threshold, the report explains the specific problem, provides rewritten text you can paste directly into LinkedIn, and lists the exact navigation steps to make the change. A sample output for a profile scoring 61/100 looks like this: headline scored 5/10 because it's the default "Software Engineer at TechCorp" format that matches zero keyword filters. The recommendation provides the new headline with pipe-separated skills and domain. The steps tell you which pencil icon to click.
There's also a manual checklist for six items that can't be verified from the CSV export (profile photo, cover photo, custom URL, featured section, Open to Work settings, visibility). For each one, the report tells you where to find the setting and what to look for.
Why structured scoring matters
The real value of your LinkedIn score isn't the number. It's the prioritization. If your headline is broken and your About section is empty, those two fixes matter more than everything else combined. The weighted scoring surfaces that automatically. You spend 30 minutes on the highest-impact changes instead of randomly tweaking whatever catches your eye.
It also catches things people miss entirely. An "SDE II" title that only makes sense inside Amazon. Three soft skills taking up slots where hard skills should go. A gap in posting activity that makes LinkedIn deprioritize your profile in search results. These are not things most people notice about their own profile because they're not thinking like a recruiter running a search filter.
Using it
The tool is packaged as an AI skill, which means it works as a structured instruction set for any LLM. In Claude Code, you clone the repo and register it as a skill. In Claude.ai Projects, you upload the skill file as project knowledge. For any other tool that accepts context files, you upload the scoring rubric and checklist alongside your CSV export and ask for an audit.
The repo is on GitHub: linkedin-profile-score
The scoring criteria are Nicole Barra's expertise made executable. I didn't invent what makes a good profile. I just thought it would be useful to have a starting point before asking someone for feedback.