[Feature] - Add agent auth support in rogue-tui

[yuval-qf]

Describe the solution you'd like
Add agent auth support to rogue-rui.
We currently support these authentication mehods: https://github.com/qualifire-dev/rogue/blob/main/sdks/python/rogue_sdk/types.py#L22
Those should be also implemented in the tui.

Note: we only need to fetch the auth data from the user and pass it to the rogue server. the server will handle the rest.

[Freeky7819]

Hi, we already prepared a minimal working prototype aligned with types.py (no_auth, api_key, bearer_token, basic).
The TUI wizard collects the agent credentials and outputs a valid agent_config JSON fragment.

Implementation is ready to open as a PR (tested locally).

Before that, could you please confirm:

Which endpoint should be used for auth test (/whoami or /ping)?

Should BASIC auth be sent base64(user:pass) or plain?

Is local config persistence acceptable, or do you prefer in-memory only?

If you confirme it, we can open a clean PR (approx. 3 small Python modules + README).

[drorIvry]

Hi @Freeky7819 I think this issue refers to the Auth options of the protocols

Meaning that the evaluator agent would be able to talk to other agents that needs some sort of Auth as part of the protocol https://modelcontextprotocol.io/specification/draft/basic/authorization for MCP

And https://a2aprotocol.ai/docs/guide/a2a-samples-hello-world#authenticated-endpoints for a2a

[Freeky7819]

Prototype repo ready:
👉 github.com/Freeky7819/rogue-auth-bridge

It implements MCP + A2A protocol authentication headers for evaluator agents,
tested locally with bearer tokens and signed passports.

[drorIvry]

Hi @Freeky7819 we're looking for active contribution to the repo, please refer to the CONTRIBUTING.md for more details on how to contribute code