prototype: support plain a2a via adding images

[tomkis]

Summary

Linked Issues

Documentation

• No Docs Needed:

If this PR adds new feature or changes existing. Make sure documentation is adjusted accordingly. If the docs is not needed, please explain why.

[gemini-code-assist]

Summary of Changes

Hello @tomkis, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly enhances the system's ability to onboard and manage agents by introducing a dynamic method for retrieving agent metadata. It moves beyond relying solely on Docker image labels, allowing the system to deploy a temporary, isolated environment to query an agent's capabilities directly from its running container. This change improves flexibility and compatibility for integrating diverse agent images into the platform.

Highlights

• Direct Agent Card Retrieval: Implemented a new mechanism to fetch AgentCard information directly from a running Docker image. If the expected 'beeai.dev.agent.json' label is missing, a temporary Kubernetes 'probe' deployment is created to extract the card via a well-known HTTP endpoint.
• Custom Exception for Missing Labels: Introduced a dedicated MissingAgentCardLabelError to specifically handle scenarios where the crucial 'beeai.dev.agent.json' label is absent from a Docker image, providing clearer error handling.
• Kubernetes Probe Deployment Management: Added comprehensive functionality within the Kubernetes deployment manager to create, manage, and tear down temporary Kubernetes services, secrets, and deployments specifically for probing Docker images to retrieve agent metadata.
• Agent Card Enrichment: AgentCards retrieved directly from containers are now automatically enriched with a default AGENT_DETAIL_EXTENSION_URI if this extension is not already present, providing consistent metadata.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces support for resolving agent cards directly from container images, which is a significant enhancement for plain A2A (Agent-to-Agent) functionality. The changes involve adding a new custom exception for missing agent card labels, implementing Kubernetes probe deployments to fetch agent cards from running containers, and integrating this logic into the provider creation and patching workflows. The code is generally well-structured and includes necessary error handling and retry mechanisms. I've identified a few areas for improvement related to error logging during cleanup, timeout configurability, and naming consistency for interface methods.

[gemini-code-assist]

Similar to the comment in provider_deployment_manager.py, suppressing all exceptions during the cleanup of probe deployments can hide important issues. It's recommended to log any exceptions that occur during delete_probe_deployment at a warning level. This ensures that cleanup failures are recorded for debugging purposes without preventing the primary error from being raised.

        finally:
            with suppress(Exception):
                await self._deployment_manager.delete_probe_deployment(probe_id=probe_id)

[jezekra1]

Probably you want to inherit PlatformError and use 4xx status code here

[jezekra1]

should be configurable - discussed offline

[jezekra1]

This can potentially take minutes (includes downloading image ~100MB), consider a separate asynchronous API with polling for unsupported images.

This is how it can work in the CLI:

1. POST /providers -> Missing card label error
2. POST /providers/discovery
3. GET /providers/discovery/{id} # Poll status

You'd need to add new table for discovery jobs and procrastinate worker to handle this async