AutoAgent · Issues· 67 open

Proposal: pluggable vector memory backend and optional Milvus support

AutoAgent publishes an unauthenticated command-execution TCP server (tcp_server.py) on host 0.0.0.0, running attacker-supplied commands as root inside a container that mounts the host workspace

OSError("handle is closed") raised while browser environment is being set

Security: LLM can create and execute arbitrary Python code via create_tool/run_tool

Security: Command injection via unsanitized input in tcp_server.py and local_env.py

Chart Library as a financial tool for AutoAgent

Feature Request: Integrate Azure AI Foundry Endpoint for LLM Integration

args[__CTX_VARS_NAME__] = context_variables TypeError: 'str' object does not support item assignment

How can I use the coding agent to run the code?

Browser Environment fails to initialize on macOS with Chromium headless mode

Repeatedly install autoagent when running

sub_agent_info = [agent_dict[sub_agent["name"]] for sub_agent in sub_agents] KeyError: 'xxx Agent'

Linux Ubuntu Failed building wheel for greenlet, while successfully built AutoAgent but unable to call from Terminal

Exception: Failed to clone the repository. The error is: None

Г