crawl4ai
crawl4ai copied to clipboard
[Bug]: endpoint /token does not require credentials
crawl4ai version
0.7.6
Expected Behavior
When running crawl4ai from the docker image with a security config of
security: enabled: true jwt_enabled: true I would expect it is required to present a secret to the endpoint POST /token in order to obtain a temporary access token.
Current Behavior
The endpoint POST /token can be called without any credentials and it still returns a valid token that can subsequently be used to call protected endpoints such as POST /crawl. This renders the security useless in my understanding, because anyone with network access to the API alone can use it.
In previous versions we had CRAWL4AI_API_TOKEN to protect the API from unauthorised usage. But sadly, this simple but effective mechanism has been removed from later versions.
Is this reproducible?
Yes
Inputs Causing the Bug
Steps to Reproduce
Build the Docker image using a config.yml with
security:
enabled: true
jwt_enabled: true
Run the Docker image
Go to http://127.0.0.1:11235/docs#/default/get_token_token_post
Obtain a token
Use the token to call POST /crawl
Code snippets
deploy/docker/server.py at https://github.com/unclecode/crawl4ai/blob/main/deploy/docker/server.py#L288 does not require any credentials to obtain a valid JWT token for subsequent access to protected endpoints:
@app.post("/token")
async def get_token(req: TokenRequest):
if not verify_email_domain(req.email):
raise HTTPException(400, "Invalid email domain")
token = create_access_token({"sub": req.email})
return {"email": req.email, "access_token": token, "token_type": "bearer"}
OS
macOS
Python version
3.10
Browser
Firefox
Browser version
No response
Error logs & Screenshots (if applicable)
No response