Only 913 of 21,754 Israeli Engineers List GitHub. Source the Other 96%

If you source engineers on LinkedIn, Hiring Assistant 2 now has a big say in who reaches the top of your list, and it favors people who link apps and post. That leaves out the large group of strong engineers who never linked anything and never post, and they are the ones you should be going after.

Calcalist quoted LinkedIn saying that profiles with a connected app get up to 2.9x more recruiter messages, and that Hiring Assistant now reads posts and comments as well as profile text. That tells you where the agents at more than 20,000 companies are sending their InMails, and where they aren't.

What did LinkedIn actually claim?

LinkedIn claims that profiles with any connected app receive up to 2.9 times as many recruiter messages, and that its AI reads posts and comments alongside the profile. Both claims come from LinkedIn itself, so treat them as a direction, not a measured cause and effect.

Connected Apps is a LinkedIn feature that lets members attach a third-party account. The app then adds a short statement about their real activity to the profile. When it launched in June, GitHub was listed as "on the way." The menu now has 26 partners, including GitHub, Replit, Behance, ElevenLabs and Medium. Members can't edit these generated summaries.

Before you build a sourcing strategy on the 2.9x figure, note three caveats:

• It's a ceiling. "Up to 2.9x" is the best case LinkedIn chose to quote, not the average.
• It covers any connected app. The figure isn't specific to GitHub. A designer with Behance linked counts the same.
• It's a correlation. Members who bother to link apps are probably more active on LinkedIn generally, and activity attracts messages on its own.

The second claim matters more for sourcers. Calcalist quotes Thieberg of LinkedIn: "AI scans not only what is written in the profile, but also posts and comments." He adds that "candidates are not making enough use of comments on the platform." Hiring Assistant now treats posts and comments as ranking input, so how often someone posts affects whether recruiters ever see them.

Then there's the scale. LinkedIn says more than 20,000 companies use its agentic hiring products, naming AMD, Cisco, Ochsner Health, Palo Alto Networks and Zillow. On Nov 4, GuruFocus reported that the number had passed 25,000 within a month. That count covers all of LinkedIn's agentic products, not just Hiring Assistant seats. Even so, a lot of recruiting teams may be running the same ranking logic over the same profiles.

How does Hiring Assistant 2 choose who gets contacted?

Hiring Assistant 2 ranks candidates on four kinds of input: profile text, posts and comments, Connected Apps evidence, and receptiveness signals. Recruiters mostly go with its picks. LinkedIn says they are four times more likely to contact candidates the agent sourced than candidates found through traditional methods.

A receptiveness score is the agent's estimate of how likely a candidate is to reply. According to coverage of the launch, it takes in engagement, Open to Work details and past InMail response. So the ranking draws on:

1. Profile text. Titles, skills and descriptions, as before.
2. Posts and comments. New, and it favors people who post a lot.
3. Connected Apps statements. Activity summaries from partners like GitHub that members can't edit.
4. Receptiveness. Engagement, Open to Work status and InMail reply history.

Once the agent has a ranking, sending is cheap. It can send messages in batches of up to 25 candidates, each one using an InMail credit, and a Recruiter seat can send up to 150 messages a month. The top of the ranked list gets messaged first.

Keep one exception in mind. HeroHunt reports that since May 2026, "professional sources" have fed trusted candidate data from GitHub, patents and similar places into Hiring Assistant's evaluations. An engineer who never linked GitHub may still have some code evidence in the agent's view. What they can't have is posts, comments or reply history. So even if the agent can see their code, they still score low on receptiveness.

How many engineers show GitHub on LinkedIn at all?

Very few: in Refolk's index of professional profiles, only 913 of the 21,754 people in Israel titled Software Engineer, Senior Software Engineer or Backend Engineer list GitHub as a skill. That's 4.2%. Only 8 mention "github" in their headline.

A caveat on method: listing GitHub as a skill is a stand-in for showing GitHub on the profile. Refolk's index doesn't record whether someone has linked GitHub through Connected Apps. But it's hard to see why linking would be common in a group where under 1 in 20 even type the word.

4.2%

The employers in the two groups also look different, though the samples are small (25 profiles each), so read this as a hint only. In the full pool, the most common employers in the sample were Wiz and Eon.io, then Lemonade and Microsoft. In the GitHub-skill group, Meta led, then Intuit and ELTA Systems. If you're sourcing from a particular set of companies, don't assume their engineers advertise GitHub the same way.

Why does the agent's attention pile onto the same engineers?

The few engineers with visible signals rank near the top for every customer running a matching search, and every reply they send raises their score further. Each step in the loop makes sense by itself, but together they concentrate outreach on a small group.

Here is how it builds:

1. Around 4% of a local engineering pool shows a GitHub signal on the profile.
2. Every customer's agent ranks those engineers near the top for any matching search.
3. Recruiters are four times as likely to contact the people the agent picked.
4. The agent sends in batches of up to 25, so working through the top of the list takes little effort.
5. The engineers who reply build InMail response history. That raises their receptiveness score, which keeps them at the top next time.

For quiet engineers, the same loop runs the other way. They aren't contacted, so they never build response history, so their receptiveness score stays low, so they aren't contacted. Being the least-contacted group says more about how the ranking works than about whether these engineers would consider a move.

This is why passive software engineer sourcing pays off right now. The visible 4% are likely getting batch outreach from many companies at once. The other 96% are competing for far less attention.

Why does a "github" keyword search mislead?

Searching for "github" on LinkedIn mostly finds where people work, not what they build. Of the 8 Israeli engineers in Refolk's index with "github" in their headline, 2 work at GitHub the company, and the rest are at large platforms such as Microsoft, Google, JFrog and Palo Alto Networks.

A Boolean "github" search in LinkedIn Recruiter brings back people who work near GitHub, not people who commit to it. The profile text was never where the code evidence lived.

If you want to find GitHub developers on LinkedIn, start at the other end. Begin on GitHub with the language, repositories and location you need, then match each person to their LinkedIn profile to check their current employer and level. That matching step is the slow part, and Refolk is built for it: you describe the engineer in plain English (the language, the kind of repos, the city) and get back a ranked shortlist with their GitHub, LinkedIn and open-web profiles already linked together.

How do you source engineers the agent ranks low?

Start with code, then find contact details, then use LinkedIn only to confirm. That way posting habits and Connected Apps never decide who you find, which is how you reach engineers who aren't active on LinkedIn.

1. Define the role in terms of code. Pick the language, frameworks and repositories or organizations a strong candidate would contribute to, and check for recent commit activity.
2. Search GitHub first. Filter users by location and language, and pull contributor lists from the repos you named.
3. Find contact details. Look for a public email address or a linked LinkedIn URL on the GitHub profile.
4. Check LinkedIn for facts only. Confirm the current employer, title and tenure. Don't screen people out because they don't post.
5. Contact them outside InMail where you can. Email doesn't compete with the 25-person batches, and it doesn't use up a seat's 150 monthly messages.
6. Mention something they built. Refer to a specific PR, repo or commit so the message reads like it came from a person who looked, not from an agent batch.

Step 3 works more often than most recruiters expect. StarHunt's 2026 study of more than 330,000 GitHub profiles found that 39% publish a public email address or a LinkedIn profile.

61%

So an engineer who hasn't linked GitHub can still be reached. For most active developers whose location you know, the contact details are already public. The hard part is connecting a GitHub handle to the right person, employer and channel, and doing that at scale.

A shortlist still isn't a hire. Someone has to screen each candidate and turn a cold message into a conversation, while agents at more than 25,000 companies keep working the visible pool. For a hard-to-fill backend or infrastructure seat, that gap is where the time goes.

Should you tell engineers to comment more?

No. Comment volume is easy to fake, and LinkedIn is already trying to police it. A commit history is much harder to fake, so it's the better signal for recruiters and a better use of an engineer's time.

Thieberg's advice to candidates was to make more use of comments. But LinkedIn also says it can identify 94% of AI-driven activity on the platform. If the company needs a detector, the signal is already being gamed. An engineer who starts producing comments to climb Hiring Assistant's ranking adds noise to the very input recruiters are being told to trust.

What this means in practice:

• For recruiters: give code evidence (commits, reviews, maintained repos) more weight than comment activity, and treat a quiet profile as a reason to go check GitHub.
• For engineering leaders: expect your strongest quiet engineers to get fewer recruiter messages than your loudest. That's good for retention until a sourcer starts from GitHub.
• For candidates: linking GitHub through Connected Apps is a reasonable, honest signal. Commenting just to be seen probably isn't worth the effort.

The engineers worth hiring next quarter are mostly not the ones getting batch InMails this week.

FAQ

Does linking GitHub on LinkedIn get engineers 2.9x more messages?

Not necessarily. LinkedIn's figure is "up to" 2.9x and applies to any connected app, not GitHub specifically. It's also LinkedIn's own number and doesn't separate the effect of linking from the fact that people who link apps tend to be more active. Treat it as evidence that the agent favors visible signals, not as a promise.

Can Hiring Assistant see an engineer's GitHub if they never linked it?

Possibly, in part. HeroHunt reports that since May 2026, "professional sources" have brought trusted data from GitHub, patents and similar places into Hiring Assistant's evaluations. But an unlinked, quiet engineer still has no posts, comments or InMail reply history, so their receptiveness score stays low even if some of their code is visible.

What's the fastest way to find GitHub developers on LinkedIn?

Start on GitHub, not LinkedIn. Search by language, repositories and location, then match each handle to a LinkedIn profile to confirm employer and level. Searching LinkedIn for "github" mostly brings back employees of GitHub and large platforms: 2 of the 8 Israeli headline matches in Refolk's index work at GitHub the company.

Are quiet engineers less likely to accept an offer?

Nothing I've seen shows they are. Their low receptiveness scores come partly from the scoring itself: they aren't contacted, so they never build reply history, so they keep ranking low. Because they hear from fewer recruiters, a well-researched message that refers to their actual code faces less competition.