galah
galah copied to clipboard
Galah: LLM-powered web honeypot. Wasting attackers' time with faker-than-ever HTTP responses!
Implement a dynamic, rule-based approach to gain more control over response generation. With this approach, we can further reduce OpenAI API costs and enhance the accuracy of generated responses. For...
If, for example, galah is returning a redirect response, it should return the appropriate return code (bonus: make return codes other than 200 optional). Current behavior: ``` curl -i http://localhost:8080/wp-admin/...
Create a simple signature file for activity tags. We can initially generate these tags using Suricata rules and analysis of received requests in our current dataset. This simple tagging will...
Store unique requests in a cache and add a tag to event logs for new or unseen requests. This will help identify and track previously unencountered activity.
Support the generation of non-text content types such as `application/octet-stream`, `application/pdf`, `application/zip`, `image/jpeg`. This feature will broaden the range of outputs, enabling more versatile use cases. For instance, it can...
Log JA4+ network fingerprints (JA4 and JA4TCP?): https://github.com/FoxIO-LLC/ja4
when i use the parameter -u to provied my open AI server it just does't seem to use that and sends the reqeust to openAI servers. which results in invalid...