browserpilot

browserpilot copied to clipboard

Enable OpenAI other compatible APIs

7

Two changes: - One added environmental variable to client setup to enable openAI compatible endpoints `OPENAI_API_BASE_URL` as environmental variable. - added strip command to OPENAIKey - Some weirdness happening with...

rapatel0

compatibility with GPT-4-Turbo models + a new instruction

2

Since `gpt-4-turbo-preview` and `gpt-4-1106-preview` append `"```"` to the response, I removed the `stop=` parameter which was useless and added a regex to fetch the python code. I also add a...

agn-7

Vulnerability Report: Code Injection Vulnerability in gpt_selenium_agent.py due to Flawed _check_danger and exec() Execution in __step_through_instructions

2

### Description The `__step_through_instructions` method in the `gpt_selenium_agent.py` file is vulnerable to CWE - 94: Code Injection. This method uses the `exec()` function to execute the `action` code, which is...

ybdesire

### Description The code snippet `__run_compiled_instructions` within the `GPTSeleniumAgent` class is vulnerable to a CWE - 94: Code Injection vulnerability. The function uses the `exec()` function to execute the `instructions`...

ybdesire

Add flag to disable is_dangerous import checking

10

I'd like the ability to disable the is dangerous check as needed through a flag: https://github.com/handrew/browserpilot/blob/b414a3c78b40df8e4aaacc022176b6fbfdc8f1e9/browserpilot/agents/gpt_selenium_agent.py#L149 My use case returns json, and generates an import statement that forces the code...

jfelten

Hi @handrew, First of all, thank you for your great work on browserpilot — it's a really useful project for AI-driven browser automation. I’d like to suggest adding optional support...

TomX99