Unable to run the Scan

[xalgord]

What happened?

I am using Gemini API and 2.5 flash model, but after hitting execute, it reopens the same window again. Can someone tell me why my scan is not running? I have even added API to envionment variables.

How to reproduce

• Install enterprise stack
• Do configuration, add api etc.
• run the scan on target.

How are you running it?

React UI (npm start)

Provider and Model

gemini/gemini-2.5-flash

Operating System

windows 11 - WSL

Python Version

3.13.7

Node Version (if using React UI)

v20.19.2

Cyber-AutoAgent Version

0.1.3

Error logs

[xalgord]

Also tried in python evironment and got this error:

18:03:58 ERROR [PythonExecutionServiceAdapter] Python execution failed:
Error: Python process exited with code 1

Solution: Check stderr output below for details

cwd=/root/Cyber-AutoAgent pythonPath=/root/Cyber-AutoAgent/.venv/bin/python
--- stderr (last 12 lines) ---
Invalid CYBER_AGENT_TEMPERATURE=null - ignoring
Unknown embedding model 'gemini/gemini-2.5-flash', dimensions not in lookup table. Attempting to infer from model name or defaulting to 1536.
Could not infer dimensions for 'gemini/gemini-2.5-flash', defaulting to 1536. If this is incorrect, the FAISS index will fail to load.
-------------------
Stack: Error: Python process exited with code 1

Solution: Check stderr output below for details

cwd=/root/Cyber-AutoAgent pythonPath=/root/Cyber-AutoAgent/.venv/bin/python
--- stderr (last 12 lines) ---
Invalid CYBER_AGENT_TEMPERATURE=null - ignoring
Unknown embedding model 'gemini/gemini-2.5-flash', dimensions not in lookup table. Attempting to infer from model name or defaulting to 1536.
Could not infer dimensions for 'gemini/gemini-2.5-flash', defaulting to 1536. If this is incorrect, the FAISS index will fail to load.
-------------------
    at sanitizeEntry (file:///root/Cyber-AutoAgent/src/modules/interfaces/react/dist/utils/logger.js:63:35)
    at Logger.log (file:///root/Cyber-AutoAgent/src/modules/interfaces/react/dist/utils/logger.js:143:27)
    at Logger.error (file:///root/Cyber-AutoAgent/src/modules/interfaces/react/dist/utils/logger.js:114:14)
    at file:///root/Cyber-AutoAgent/src/modules/interfaces/react/dist/services/PythonExecutionServiceAdapter.js:218:20
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)









○ local-cli                                                                                   gemini/gemini-2.5-flash | 0 tokens | $0.00 | [ESC] Kill Switch
Error: Python process exited with code 1

Solution: Check stderr output below for details

cwd=/root/Cyber-AutoAgent pythonPath=/root/Cyber-AutoAgent/.venv/bin/python
--- stderr (last 12 lines) ---
Invalid CYBER_AGENT_TEMPERATURE=null - ignoring
Unknown embedding model 'gemini/gemini-2.5-flash', dimensions not in lookup table. Attempting to infer from model name or defaulting to 1536.
Could not infer dimensions for 'gemini/gemini-2.5-flash', defaulting to 1536. If this is incorrect, the FAISS index will fail to load.
-------------------
    at ChildProcess.<anonymous> (file:///root/Cyber-AutoAgent/src/modules/interfaces/react/dist/services/PythonExecutionService.js:886:39)
    at ChildProcess.emit (node:events:524:28)
    at ChildProcess._handle.onexit (node:internal/child_process:293:12)

[westonbrown]

Gemini2.5 flash is not an embedding model.

You need to configure with an embedding model like titan embedding on bedrock or another provider you can also use local text embeddings from huggingface.

This is a required model to support long term memory.

[xalgord]

actually i only have google cloud credits, is there any way to use only google api? like i've heard of genai, but what will be the variable value to configure it?

[xalgord]

and can someone please share a guide on AWS setup for this?

[westonbrown]

@xalgord yes you can use google studio or vertex ai apis through litellm on the release 0.1.4 branch.

This is not a bug but a configuration issue. Please review in the release branch how to configure in the deployment doc.

Thanks!