LinkedIn detection no longer works due to blocked requests (Request Denied – HTTP 999)

[aryanj10]

Installation method

PyPI (via pip)

Package version

Sherlock v0.16.0

Description

LinkedIn username detection currently fails. When running:

sherlock aryanj10 --site "LinkedIn"

the output shows “0 results,” even though the account exists (e.g. https://www.linkedin.com/in/aryanj10/
).

Observed behavior:
LinkedIn returns an HTTP 999 Request Denied for all automated requests.
Example:

curl -I https://www.linkedin.com/in/aryanj10/
HTTP/1.1 999 Request denied

Expected behavior:
Sherlock should detect the account or, if detection is impossible, mark LinkedIn as deprecated/unusable in data.json.

Suggested fix:
Remove LinkedIn or disable it by default since it consistently blocks requests with HTTP 999 and cannot be reliably queried even with headers or Tor.

Steps to reproduce

sherlock aryanj10 --site "LinkedIn"

Additional information

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[Gauravpadam]

What would be a plausible fix here @ppfeister @aryanj10?
Could we mark linkedin as deprecated until we figure out a way to query it?

[KingSlayer-KS]

Would you prefer I handle LinkedIn by adding a site-specific header patch inside SitesInformation, or just disable it in data.json since it’s consistently returning HTTP 999?

[ppfeister]

Busy weekend so unable to opine on fixes right this second ---- but on the header patch.... no site-specific modifications should be made to the core of the project. However, there is a headers attribute allowed in the schema for data.json. A few sites already use it (unless they've been removed at some point), so there are some examples of its implementation available.

[sonianuj287]

Hi @aryanj10 , what you prefer to handle this issue.
possible options I can think of -

1. Mark LinkedIn support as deprecated / disabled by default, with clear messaging
2. Implement a fallback / workaround, e.g. using different headers, user-agent, or proxy, if feasible
3. Add detection logic to catch HTTP 999 and output a warning, not a “0 results” silently