indymedia
/
epicyon
mirror of https://gitlab.com/bashrc2/epicyon
7
4
Fork 1
Code Issues 14 Releases Wiki Activity

Log before setting headers

main
Bob Mottram 2024-08-18 10:44:51 +01:00
parent 2f41ccc0ce
commit 515b9243cc
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
daemon_get.py
View File

@ -263,13 +263,13 @@ def daemon_http_get(self) -> None:
# oai-host-hash requests come from Microsoft Corporation, # oai-host-hash requests come from Microsoft Corporation,
# which has a long term partnership with OpenAI # which has a long term partnership with OpenAI
if 'oai-host-hash' in self.headers: if 'oai-host-hash' in self.headers:
print('GET HTTP LLM scraper poisoned: ' + str(self.headers))
msg = html_poisoned(self.server.dictionary) msg = html_poisoned(self.server.dictionary)
msg = msg.encode('utf-8') msg = msg.encode('utf-8')
msglen = len(msg) msglen = len(msg)
set_headers(self, 'text/html', msglen, set_headers(self, 'text/html', msglen,
'', calling_domain, False) '', calling_domain, False)
write2(self, msg) write2(self, msg)
print('GET HTTP LLM scraper poisoned: ' + str(self.headers))
return return
# replace invalid .well-known path, prior to checking for suspicious paths # replace invalid .well-known path, prior to checking for suspicious paths
@ -336,13 +336,13 @@ def daemon_http_get(self) -> None:
if block: if block:
if llm: if llm:
# if this is an LLM crawler then feed it some trash # if this is an LLM crawler then feed it some trash
print('GET HTTP LLM scraper poisoned: ' + str(self.headers))
msg = html_poisoned(self.server.dictionary) msg = html_poisoned(self.server.dictionary)
msg = msg.encode('utf-8') msg = msg.encode('utf-8')
msglen = len(msg) msglen = len(msg)
set_headers(self, 'text/html', msglen, set_headers(self, 'text/html', msglen,
'', calling_domain, False) '', calling_domain, False)
write2(self, msg) write2(self, msg)
print('GET HTTP LLM scraper poisoned: ' + str(self.headers))
return return
http_400(self) http_400(self)
return return