Adam Swanda
Adam Swanda
Pastebin Pro offers an API that is meant for scraping specifically. It's only available to paid users, though having the option of the standard scraping versus scraping API would be...
It'd be great if the regex patterns were moved to a config file, and users could also specify direct keywords they want to find, so you don't need to edit...
When running `python3 ddom.py -s malshare --download` the module will fail with an error/debug message of `can only concatenate str (not "bytes") to str`. For the life of me I...
Review logging across the application and standardize to a common format that's conducive to key-value extraction. I'm thinking two main loggers: * Application logs (database, loading and running scanners, etc)...
Utilize Flask's Blueprint feature to separate different routes into their own modules. This allows grouping related routes and keeps the main API server application cleaner.
Now that Vigil can be imported as a library, it is a lot easier to call the different scanners and interact with canary tokens. This allows users to define custom...
Research and implement detection strategies for image prompt injection. OpenAI's GPT vision API is now available so I can do some better testing. https://platform.openai.com/docs/guides/vision
Migrate the API server use FastAPI instead of Flask. I'm more familiar with Flask but after reading about fastapi a bit, it seems preferable for APIs. The Streamlit UI won't...
Vigil should have a helper utility to assist users in evaluating datasets against the detection mechanisms and return true positives, false positives, false negatives, accuracy score, etc. I'm thinking a...
**Add support for Cohere embeddings** - Create datasets for Huggingface - Update vectordb scanner to use Cohere as custom embedding function - Update config parser to load Cohere model choice...